The Link n°120

LSR: What we put forward to Commissioner Serafin

On 28 September 2026, several trade union and staff representatives were invited by Commissioner Serafin to discuss the modernisation of our institution as part of the Large Scale Review (LSR). They were given the opportunity to voice their views and concerns. Our President shared his thoughts on the work undertaken and the findings of the various working groups set up for this purpose. The following remarks are in line with this approach, aiming to explain the purpose of this process, the challenges it presents and how it can strengthen our collective action.

First, Commissioner, I would like to thank you for allowing us to participate in this major exercise to overhaul our organizational practices — the LSR (Large Scale Review). For us, this exercise is driven first by severe budgetary constraints at a time when our Institution must undertake significant initiatives essential to the survival of the European Union. However, it is also important because the Commission must constantly improve — wherever possible and at an appropriate pace.

Today’s dialogue marks an essential step, but one that must be continued and deepened as soon as the group led by Catherine Day presents its recommendations on the matter. More importantly, our dialogue must continue when your own positions are made known to us, and indeed during their implementation. After all, as is said, the devil is in the details?

For the moment — apart from the proposals already put forward by our organization and the joint proposals from the trade union alliance, we wish to share some thoughts regarding the findings of the various working groups established under this LSR. We also wish to address issues that are not currently part of this initiative.

The LSR omits discussion on the social dialogue and how it might be improved. This is a matter of common interest for which you bear [BARE] just as much responsibility as we do. This dialogue is part of our shared values; its implementation is framed by the Treaties and promoted by the “European Pillar of Social Rights,” which our President is committed to supporting.

A second missing issue concerns the EEAS. At the time of its creation, our organization advocated for its integration into our Commission, where it belongs. We argued that this would make its work more efficient, avoid duplication, ensure better coordination between the external dimensions of our various policies, and finally, offer better career prospects for staff. Above all, its integration into our Commission is essential because the Commission is the only institutional setting that enables the formulation and expression of a common European interest — which is, after all, our core business.

We are aware that sectoral, corporatist, and nationalist interests still stand in the way today. However, the current political landscape, the time that has elapsed since the service was established, and its track record all support our position.

Let us now turn to the LSR as it currently stands — and recognising that not everything has been completely revealed.

My first concerns recruitment, integration into the services, and career progression. Recruitment needs to be diversified and improved to take account of candidates’ interests — which is currently not the case — and to better assess their practical capabilities. These capabilities need to be better evaluated and valued.

We therefore need a combination of improved external competitions, internal competitions, and “Young Professional” schemes. Career progression should provide opportunities for staff to develop themselves, their competences and the quality of their work, rather than simply multiplying hierarchical levels.

Above all, career paths must be diverse and merit based. However, career progression must strike a balance between a logic of meritocracy (letting the best succeed) and a logic of cohesion (ensuring everyone has the right to career development).

My second point concerns the formulation of an ambitious policy for contract and temporary staff. These staff members will remain essential to our Institutions and deserve career paths that, for the best among them, lead to permanent official status. They viewed the Large-Scale Review as an opportunity to improve their situation. Indeed, they mobilized for this reason.  

We have a moral and political obligation to listen to their concerns. It is also in our interest to do so, including from a budgetary perspective.

The third point concerns agencies and first executive. In our view, they should be brought in-house within the Commission as offices.

This possibility is being considered as part of the ongoing discussions regarding the separation of political, implementation and support functions. This split could lead to the creation of new executive structures which we hope will become offices.

Integrating them would make work more efficient and cost-effective and would benefit the staff. Here, too, we encounter resistance from certain groups and conservative corporate interests — resistance, which we consider mostly unjustified. We must pay particular attention to attempts at outsourcing; let us not forget the lessons of the technical assistance offices of the past.

As for decentralized agencies, we must first strengthen governance and our presence. As you know, we advocate for the Commission take direct responsibility for several services.

Our representatives on their management boards must pay greater attention to the implementation of the Staff Regulations and to compliance with the Treaties, particularly concerning the social dialogue.

Ultimately, the most critical issue remains the excessive and uncontrolled increase in the intensity of work in a context of deteriorating budgetary conditions, which poses both organizational and human risks. Doing more and doing it better with fewer resources may sound modern, but it is not always desirable— and, above all, not always possible.

This increase in workload is primarily due to the hybrid nature of our work, which boosts productivity. This is not without risks for staff and for the quality of the work itself. This intensification is set to increase further with the introduction and widespread adoption of AI (artificial intelligence). Finally, the matrix based organizational structure advocated in the LSR risks exacerbating this process. The challenges are immense regarding training, the reorganization of departments, and staff well-being and health — and, as the documents highlight, – notably the risk of “cognitive atrophy” caused by the use of artificial intelligence.

While AI offers nothing but advantages for certain tasks — such as salary and pension payments, health expenditure reimbursements, translation, documentation, and budget execution monitoring — its widespread adoption poses significant challenges regarding training, reorganization, and working practices. This is why it is essential to evaluate it on an ongoing basis, and to continue our social dialogue.

Thank you for your attention!

Georges Vlandas,

President of U4U

---

---

Should certain functions of the European Parliament be transferred to an office? Yes! It is time to move forward.

For years, U4U has campaigned for one clear objective: to enable certain professions within the European Parliament, particularly security staff, to be attached to a European Parliament office.

Why?

This is not merely an administrative change, but a matter of recognition and career prospects.

Currently, office-based contract staff (regardless of their function group, I, II, III or IV) may be offered a contract under Article 3a of the Conditions of Employment of Other Servants of the European Union. In practice, this means that, after a second renewal, they may be offered a permanent contract.

Currently, this is not possible for Group I contract staff employed on permanent contracts at the European Parliament. They cannot be reclassified into a higher function group without losing the benefits of their permanent contract. However, solutions already exist.

The European Financial Regulation clearly defines an office as an administrative structure created to carry out specific horizontal tasks on behalf of the European institutions.

These offices have been in operation for a long time. The Office for Infrastructure and Logistics in Brussels (OIB), the Office for Infrastructure and Logistics in Luxembourg (OIL), the Paymaster’s Office (PMO), and the European Personnel Selection Office (EPSO) are just a few examples.

So, why not take it a step further?

Technical functions and security roles could easily be assigned to an existing office, such as the OIB or the OIL, or to a specific office within the European Parliament that has been set up to carry out these essential tasks.

This would be a significant step forward:

• To finally open up genuine career prospects for the relevant staff;
• To enable progression to higher function groups where skills and experience warrant it;
• To fully recognise the indispensable role played by these staff members in serving the European institutions every day.

This reform is by no means utopian. It is legally feasible, administratively realistic, and politically necessary.

---

The Assembly of Agency Staff Committees (AASC): Building the Future of Our Agencies Today

Both the elections for the new secretariat of the Assembly of Staff Committees of the Agencies (AASC) and the 51(st) meeting of this body are taking place from 9 to 11 November 2026. To mark this occasion, trade union organisations have been invited to participate in a round-table discussion on working conditions within European agencies. In this context, our president, Georges Vlandas, presented our organisation’s trade union perspective on the main concerns regarding working conditions in European agencies:

Dear Colleagues,

I am pleased to be here with you for this, the AASC’s annual meeting.

Based on the assessment of your activities, the purpose of your meeting is to present some perspectives for the future and to elect the members of your secretariat.

Allow me to share with you our 5 main concerns:

1. Your working conditions depend, above all, on the Union’s budget, specifically its total amount. Staff should be fully cognisant of this. The Commission’s trade unions have just sent a message to the presidents of the European Institutions asking them to publicly advocate for an increased budget and to explain to citizens the reasons behind such a decision. A reduction in the EU budget would lead us collectively into a dead-end and would inevitably result in a further reform of the Staff Regulations. This would therefore be detrimental to both European society and the European civil service.

Together with other trade unions, we have called on our leaders to bring the budget debate into the public arena, so as to involve the citizens whom we ultimately serve. The proposed budget for the future is inadequate; it will not enable us to meet the challenges we face. It will not allow the European civil service to function properly. Yet solutions do exist; they are well known – as well-documented in the Draghi report. The fact that they are not being implemented illustrates the severity of the crisis we are facing—a crisis to which our work aims to provide solutions. It is for these reasons that I urge you to formally support and adopt the text of the Commission’s Staff Trade Unions.

• We also need to strengthen our social dialogue. Nothing of value can be achieved without involving staff. This dialogue, in which you are involved, must be improved. Firstly, the AASC must become a joint staff committee representing all staff committees at a central level. Such a committee, with a statutory status, will be given greater recognition. Consultation with it would be mandatory and its views would carry greater weight.

• However, a stronger staff committee is not enough; it is necessary to strengthen trade union representation in the decentralised agencies and the executive agencies, and thus further boost the social dialogue. Not all agency directors are in favour of such an approach. Yet social dialogue, including the procedures governing it, is enshrined in the Treaties. It is widely practised in the Member States and, in an exemplary manner, at the Commission. This is why I propose that your assembly adopts a resolution requiring the creation of trade union sections in each agency.

• Furthermore, regarding the governance of decentralised agencies, as the LSR’s work indicates, it will be necessary to ensure the development of shared services between agencies, in coordination with the relevant Commission departments, for the management of issues such as bullying and harassment in the workplace, mediation, disciplinary investigations and the organisation of staff mobility, in order to guarantee both the independence of the decisions taken and their impartial implementation. Some agencies are entering into agreements with the Commission in this regard in relation to bullying and harassment. This trend must be continued and expanded and should even be extended to the other areas mentioned above. And, if you agree, your assembly could make it a formal requirement.

• Furthermore, certain decentralised agencies – just like other decentralised agencies, for that matter – employ civil servants. We are in favour of maintaining, or even increasing, their numbers within and across agencies, to ensure greater independence, mobility and broader career prospects for staff.

In addition, the dispersed, clientelistic distribution of decentralised agencies across the EU’s territories and within individual countries limits staff’s opportunities for mobility and their children’s education. The perfect example is provided by Spain, where this dispersed distribution of agencies and institutional services across no fewer than seven locations prevents the creation of a local labour market that benefits staff and provides educational opportunities for their children. This makes the various locations less attractive to non-national staff. We cannot easily reverse decisions of the past but we learn to avoid repeating past mistakes.

Finally, as regards the executive agencies, their integration into the Commission’s departments – a proposal put forward by the Central Staff Committee and the Commission’s trade unions – which would in our view have been beneficial for staff, and for the Institution’s functioning, is unlikely to materialise. On the other hand, the establishment of shared services (human resources, finance, training, mobility, etc.), or even mergers (resulting in a reduction in the number of agencies), is possible. In any case, any reorganisation should be preceded by a social dialogue involving staff committees and trade unions. Where appropriate, consultation forums should be set up.

To conclude my remarks, I wish you every success with your work. You can count on our full and active support.

Kind regards,

Georges Vlandas,

President of U4U

---

Recommendations for a policy to prevent and combat harassment based on an integrated approach

Union for Unity calls for the anti-harassment policy to be managed as an interconnected chain, with every link being equally strong. This includes primary, secondary and tertiary prevention measures, support for victims (including assistance with returning to work) and enforcement measures. This paper sets out our expectations for each of these areas, particularly emphasising the systemic dimension and the effectiveness of enforcement measures. It also proposes the immediate recognition of any person who reports being affected, regardless of the outcome of the investigation, and the incorporation of four cross-cutting priorities: the treatment of reprisals as a standalone offence; neurodiversity; the isolation of staff in regional offices or delegations; precarious employment status; and a systemic analysis of the organisational causes of harassment.

1. Introduction

Union for Unity commends the work carried out since the post of Chief Confidential Counsellor (CCC) was created, particularly the significant number of colleagues who have sought its services. This note is part of a shared objective: to ensure that all colleagues at the Commission, its agencies and its delegations can work in a safe, dignified and respectful environment. We submit these recommendations in a spirit of cooperation, believing that they will contribute to the next stage in consolidating the anti-harassment framework.

From a collective perspective, in a working environment where resources are scarce and the workload is increasing, we cannot afford the harmful effects that harassment has on the targeted individual, their immediate team, and the wider working circles surrounding them. This often results in the victim’s health deteriorating, sometimes leading to long periods of absence without cover. This means more work for the immediate team or a lower standard of work.

• The chain principle: five links, one strength

A policy to prevent and combat harassment is only as strong as its weakest link. Focusing efforts on a single aspect, such as the quality of support and guidance provided to victims, without strengthening the others, will not protect staff in the long term. We urge the CCC to pay equal attention to the five areas summarised below before outlining the areas where we believe strengthening is particularly necessary in sections 3 to 7.

1 Primary prevention

→

2 Secondary prevention

→

3 Tertiary prevention

→

4 Support

→

5 Enforcement measures

1. Primary prevention — addressing work organisation and managerial culture

The aim of this component is to reduce the occurrence of harassing behaviour by addressing its organisational causes, such as workload, the clarity of roles, management style and the autonomy granted to unit managers in managing their teams.

• Incorporate a ‘psychosocial risks’ component into the assessment of any departmental reorganisation, from the outset rather than retrospectively.
• Make access to management positions and renewal of appointments conditional on an assessment that includes the working environment as perceived by the team.
• Strictly regulate the use of annual appraisals as a management tool.
  
  

For example, the Harassment Watch Network (HWN) has documented how the annual appraisal process has been used to isolate or discredit targeted staff members under the guise of a neutral administrative procedure. This shows that seemingly innocuous management tools can become a means of exerting pressure if they are not properly regulated.

• Secondary prevention: identifying early warning signs before the situation escalates
• Roll out and make compulsory ‘bystander intervention’ training, covering immediate responses to ‘micro-aggressions’, compulsory for all managerial and other staff.
• Systematically analyse the results of the staff survey at directorate-general level, providing documented feedback and implementing long-term monitoring.
• Strengthen the network of confidential advisers in terms of both numbers and ongoing training.
  
  

For example, a unit head who has been trained to spot early warning signs can defuse tension before it escalates. In contrast, a late report often comes after several months of silently deteriorating working conditions.

• Tertiary prevention: protecting the victim from the moment the first reports are made
• Establish the principle that the mobility of the accused person should be prioritised over that of the victim.
• Ensure that the accused is monitored following a change of posting to prevent the problem from simply being ‘passed on’ from one unit, department or agency to another.

For example, the systematic relocation of victims rather than the accused, a practice which is still reported in several testimonies gathered by the HWN, effectively makes the victim bear the cost of reporting the incident.

• Support for victims, including helping them to return to work

Scientific research has documented the significant and long-lasting consequences of exposure to harassment on mental and physical health, including anxiety, depressive disorders, post-traumatic stress symptoms and burnout, which can persist for several years after the events (Nielsen & Einarsen, 2012). This justifies support that is tailored to these documented effects rather than support that is limited in duration by default.

• Establish a formal protocol for returning to work following an absence due to harassment. This should involve the medical service, the line manager, and a trusted advisor chosen by the employee.
• Ensure that psychological support is provided for as long as necessary and is not limited by a fixed number of sessions.
• Ensure that career progression rights are maintained during periods of absence or transition, to avoid a double penalty.

For example, an employee returning from a long period of sick leave due to harassment may be reassigned to the same department where the problem originated, without a return-to-work protocol or a change of environment. This increases the risk of a relapse.

• The disciplinary aspect is still inadequate

The Union for Unity wishes to emphasise this aspect. For a prevention scheme to be credible to staff, proven incidents must lead to effective administrative investigations and proportionate consequences. However, the Commission has formally recognised only two cases of harassment in fifteen years, which is a stark contrast to the hundreds of requests for help that the CCC receives each year. This discrepancy fosters a sense of impunity among some perpetrators of toxic behaviour and discourages potential victims.

We would particularly like to emphasise one methodological point: the administrative inquiry should not be viewed as an exercise primarily focused on defending the institution or the rights of the person under investigation. Rather, its primary aim should be to protect the victim during the inquiry itself, not just at its conclusion.

• Set clear, publicly available time limits for each stage of the formal procedure and implement automatic follow-up in the event of delays.
• Ensure that precautionary measures to protect the victim, such as the removal of the accused person or temporary adjustments, can be taken as soon as the investigation is opened, without waiting for its conclusion.
• Ensure the systematic involvement of OLAF, as an independent authority, in formal proceedings.
• Ensure the legal robustness of investigations by implementing confidentiality measures for witness statements, establishing clear rules for data retention and adopting a balanced approach to access rights for both victims and accused individuals. This will prevent the undermining of procedures on procedural grounds, as highlighted in the opinions of the European Data Protection Supervisor (EDPS).
• Publish statistics on reports, investigations opened, cases upheld and sanctions imposed annually in anonymised and aggregated form.
• Ensure that sanctions are proportionate to the seriousness and recurrence of offences, and do not systematically apply the lowest sanctions set out in the Statute.

For example, several Members of the European Parliament (MEPs) who were found guilty of harassment received nothing more than a reprimand or a fine, which are the lowest possible sanctions. This illustrates the risk that such sanctions may have little deterrent effect. This risk is also present at the Commission, given the very low number of formally recognised cases.

• Acknowledge the affected person immediately, regardless of the outcome of the investigation

The Union for Unity would like to see a simple yet essential principle enshrined more clearly in the framework and communicated more clearly to staff: anyone who reports harassment should be recognised as someone affected by a difficult work situation from the moment the report is made, and should therefore receive immediate personal and administrative support.

However, this recognition must never be confused with the validation of allegations made against the accused person. Establishing the facts is the sole responsibility of the administrative investigation, which must be conducted in accordance with the right to be heard and the presumption of innocence of the accused. This distinction protects everyone involved. It ensures that the person making the report receives immediate and unconditional support, while also preventing this support from being mistakenly interpreted as a premature judgement on the person under investigation before the facts have been established. Recognising the hardship experienced by one person and establishing the liability of another are two distinct processes, each of which must be able to proceed according to its own logic and at its own pace.

Research in the psychology of trauma sheds light on this issue through the concept of ‘institutional betrayal’: when an institution fails to prevent or respond to harm, it exacerbates the distress experienced by the person reporting it, regardless of the facts (Smith & Freyd, 2014). This is another reason why support should not be withdrawn once the investigation has concluded.

This principle must also apply once the proceedings have concluded. Even if an investigation fails to establish that harassment has taken place within the statutory meaning due to insufficient evidence, the person who reported the situation has still experienced a difficult period at work which may have impacted their health. Support should therefore never end automatically, nor should it depend exclusively on the final legal conclusion reached by the investigation.

• Ensure that psychological support and administrative assistance are available immediately upon reporting an incident, without waiting for the investigation to begin or conclude.
• Train investigators and managers to distinguish clearly between acknowledging the situation experienced by the individual and establishing liability, to avoid any confusion that could be detrimental to either party.
• Ensure that support continues or is adapted, including psychological support, and a change of department if necessary, to avoid a toxic working environment — even if the investigation does not confirm harassment as defined by the regulations.

For example, an employee whose report did not lead to a finding of harassment due to insufficient evidence has, in most cases, nevertheless experienced a genuine deterioration in their working conditions. Maintaining their access to support beyond the closure of the case prevents them from feeling doubly penalised — first by the initial situation and then by the outcome of the proceedings.

• Retaliation is a breach in itself and should be treated as such

The Union for Unity is calling for any form of intimidation or reprisal against a colleague who wishes to lodge a complaint or give evidence in a harassment case to be explicitly classified as a separate offence in the anti-harassment decision. This classification should be independent of the outcome of the initial complaint and subject to a separate disciplinary procedure.

The risk of this happening is not theoretical: research into the reporting of workplace malpractice shows that people who report a problem are exposed to a significantly increased risk of retaliation and behaviour amounting to harassment from both management and colleagues in the period that follows (Bjørkelo, 2013).

• Introduce an explicit definition of reprisals (including, in particular, sidelining, demotion, unfavourable performance appraisals following the report, and pressure on colleagues) and classify them as a separate disciplinary offence.
• Enable an accelerated protection procedure to be initiated at the first signs of retaliation, without waiting for the investigation into the initial allegations to conclude.

For example, the widely publicised case of an EEAS official who was dismissed after reporting harassment illustrates the risk of retaliation if an incident is not treated as a disciplinary offence in its own right. The official was subjected to a disciplinary procedure lasting several years after reporting the harassment.

• Three situations of particular vulnerability to be taken into account
  • Neurodiversity

Statistically, people with neurodivergent traits (such as autism, ADHD and dyspraxia) are more likely to be subjected to harassment. This is particularly because their communication and functioning methods may be misunderstood by line managers or colleagues, leading to erroneous judgements about their behaviour or performance.

A recent clinical study confirms that a high proportion of adults with autism experience harassment, which constitutes a significant risk factor for associated psychopathological disorders in both adulthood and adolescence (Ferrigno, Cicinelli & Keller, 2022).

• Raise awareness among confidential counsellors, investigators, and managers of the potential impact of neurodiversity on vulnerability to harassment.
• Adapt investigation and support procedures to the specific needs of a neurodivergent victim without casting doubt on their credibility.

For example, repeated remarks about a neurodivergent colleague’s supposed lack of ‘social flexibility’, which is in fact linked to their cognitive functioning, may constitute discriminatory abuse if management does not identify it as such.

• Staff isolation in agencies, EU representations and delegations

By their very nature, colleagues posted to decentralised agencies or EU representations and delegations are further removed from the Commission’s decision-making centre and support network. This echoes the situation documented at Fusion for Energy (F4E) in late 2025, where there were delays in implementation and governance was deemed to lack accountability, as well as USHU’s findings regarding delegations, where mechanisms exist but are still perceived as insufficiently adapted to geographical and hierarchical distance.

A study of workers in remote locations confirms that effective access to social and institutional support significantly mitigates the impact of harassment on mental health, yet it is precisely these colleagues who are furthest removed from the usual support structures and lack such support the most (Pauksztat, Salin & Kitada, 2022).

• Ensure that agency staff have effective access to the CCC and the network of confidential counsellors, including via channels that bypass local line management.
• Incorporate the Commission’s delegated services within the remit of the CCC.
• Provide the CCC with an explicit supervisory mandate and sufficient resources across all agencies and joint undertakings, offering more than just ad hoc support in the event of a crisis.
• Make information and awareness-raising materials available in additional official languages to ensure they remain accessible to local staff in delegations who do not primarily work in French or English.

For example, a staff member at a small agency may be reluctant to report inappropriate behaviour due to a lack of a point of contact who is perceived as genuinely independent from their immediate line management.

• The precarious employment status of trainees, agency staff, temporary staff and contract staff

Those in precarious employment situations, such as trainees, agency workers, contract staff, temporary staff whose contracts are coming to an end and probationary staff, are particularly vulnerable. The fear of jeopardising the renewal of their contract or confirmation in their post may deter them from reporting inappropriate behaviour, even though they are often in a position of direct dependence on the person who may be implicated.

Research also confirms a statistical link between job insecurity and exposure to harassment from the perspective of both victims and alleged perpetrators (De Cuyper, Baillien & De Witte, 2009). Therefore, precarious employment undermines the ability to protect oneself and report incidents.

• The anti-harassment policy must explicitly guarantee robust protection and effective access to procedures for trainees, contract staff, and probationary staff, regardless of the stage of their contract.
• Provide a reporting channel that is independent of the person responsible for assessing contract renewal or job confirmation.
  
  
  
• The systemic dimension: a cross-cutting priority

The Union for Unity wants to see systemic issues permeate the entire anti-harassment policy, moving beyond the single aspect of primary prevention. How do the ways in which we organise work, define performance, distribute hierarchical power, and manage careers create conditions that are conducive to harassing behaviour emerging? This does not deny the individual responsibility of those who commit harassment but recognises that certain organisational structures may facilitate or normalise it.

This approach aligns with the ‘work environment hypothesis’, a term coined in seminal research on workplace harassment, which identifies organisational factors such as workload, role ambiguity and management style as determinants that are at least as important as the individual traits of those involved (Leymann, 1996).

• Systematically analyse the organisational causes of each confirmed incident of harassment, going beyond mere individual sanctions, in order to identify any structural factors requiring correction within the relevant department.
• Make this analysis a regular feature of the CCC’s annual activity report to monitor developments over time.

For example, a unit where the workload is too high in relation to the tasks assigned creates a breeding ground for tensions and high-pressure behaviour, even in the absence of any initial intention to cause harm.

• Monitoring, evaluation and social dialogue

Finally, the Union for Unity hopes that these recommendations will form part of a dynamic framework that is regularly reviewed in consultation with staff representative organisations and the Harassment Watch Network.

• Include a clause providing for the periodic review of the anti-harassment policy (for example, every five years), accompanied by regular staff surveys on perceptions of harassment and the effectiveness of the measures in place.
• Organise an annual structured forum between the CCC and the trade unions to share lessons learnt from cases dealt with, whilst respecting the confidentiality of those involved.

A practical example would be a statistical report shared annually with the social partners, which would make it possible to verify whether the gap between the number of reports and the number of formally recognised cases is narrowing over time.

Conclusion

The Union for Unity reaffirms its desire to contribute, in a spirit of cooperation, to consolidating a system in which every aspect — from primary and secondary prevention to support for victims and enforcement measures — is equally robust. This system must also fully incorporate the immediate and unconditional recognition of anyone affected by harassment, regardless of the outcome of the investigation. Other key aspects include fighting reprisals, neurodiversity, isolating colleagues in branch or regional offices, addressing precarious employment status, and conducting a systemic analysis of the organisational causes of harassment. We are available to discuss these recommendations in more detail with the CCC and contribute to the forthcoming review of the Commission’s anti-harassment framework alongside the Harassment Watch Network and other social partners.

___________________________

Scientific references cited

Where the scientific literature provides useful insights, this note draws on the following works:

Bjørkelo, B. (2013). Workplace bullying after whistleblowing: Future research and implications. Journal of Managerial Psychology, 28(3), 306–323.

De Cuyper, N., Baillien, E., & De Witte, H. (2009). Job insecurity, perceived employability and targets’ and perpetrators’ experiences of workplace bullying. Work & Stress, 23(3), 206–224.

Ferrigno, S., Cicinelli, G., & Keller, R. (2022). Bullying in Autism Spectrum Disorder: prevalence and consequences in adulthood. Journal of Psychopathology, 28(3), 127–134.

Leymann, H. (1996). The content and development of workplace bullying. European Journal of Work and Organisational Psychology, 5(2), 165–184.

Nielsen, M. B., & Einarsen, S. (2012). Outcomes of exposure to workplace bullying: A meta-analytic review. Work & Stress, 26(4), 309–332.

Pauksztat, B., Salin, D., & Kitada, M. (2022). Bullying behaviour and employee well-being: How do different forms of social support buffer against depression, anxiety and exhaustion? International Archives of Occupational and Environmental Health, 95(7), 1633–1644.

Smith, C. P., & Freyd, J. J. (2014). Institutional betrayal. American Psychologist, 69(6), 575–587.

---

U4U’s proposals for the upcoming housing allowance review

Almost two years after its introduction, the housing allowance paid to EU institution staff in Luxembourg has proven its worth. However, the amount has remained unchanged while housing costs have continued to rise (by 52.4 per cent compared to Brussels). U4U believes that the allowance should be increased to make the purchasing power of colleagues in Luxembourg more comparable to that of their colleagues in Brussels. Currently, the allowance is only available to staff up to grade AST5/AD5 and to contract staff up to grade GF IV.

U4U is putting forward six proposals:

1. Extend the allowance to all staff up to the equivalent of at least grade AST/AD9;
2. Increase the allowance from €352–500 to €500–750 per month;
3. Index the allowance annually;
4. Make it sustainable for at least seven years, renewable to coincide with the next MFF instead of the current four years;
5. Extend it to all trainees at institutions based in Luxembourg, as is already the case for the Commission’s ‘Blue Book’ trainees;
6. Publish the key findings of the annual evaluation (beneficiaries, categories concerned, costs and impact on recruitment).

U4U is also calling for a mid-term review clause and the allowance to be included in the next MFF budget. They are also calling for a joint approach with the other trade unions.

Details of our analysis and proposals

Almost two years after its introduction, the housing allowance has proven its worth. It has provided support to many colleagues and, to a certain extent, improved Luxembourg’s appeal as a posting location.

However, whilst the housing market has continued to rise, the housing allowance has remained unchanged.

U4U believes that the allowance’s main purpose is to achieve parity in purchasing power between staff working in Luxembourg and those working in Brussels. For the time being, this measure only applies to staff in salary categories up to AST5/AD5 (Temporary Agents AST5/AD5 and Contract Agents up to and including AC IV).

Furthermore, evidence of the critical housing situation in Luxembourg is that 80 per cent of the population is eligible for social housing[1] , subject to certain conditions. This demonstrates that the income required to find accommodation on the private market is so high that many employees could meet the criteria for housing intended for low-income households.

This is indeed why a significant proportion of staff in Luxembourg choose to live outside the country, thereby increasing their commute time. This situation is not found in Belgium or other countries that host European Union institutions.

In this context, the housing allowance should be increased and extended to consolidate the current scheme and ensure equal treatment for all staff, as well as to strengthen the effectiveness of this mechanism in the face of market changes. An increase would also help maintain and enhance Luxembourg’s attractiveness as a location.

To this end, U4U is proposing the following improvements:

1. As a first step, this allowance should be granted to all staff members up to at least salary level equivalent to grade AST/AD9. This will ensure equal treatment in terms of purchasing power for staff in Brussels and Luxembourg, and maintain Luxembourg’s greater attractiveness. Furthermore, we would like to emphasise that lower relative purchasing power in relation to housing in Luxembourg affects all staff members at all grades when compared to Brussels.
2. Similarly, we are calling for an increase in the housing allowance. As mentioned above, housing costs have continued to rise since January 2025, while the housing allowance has remained unchanged despite already being insufficient. Consequently, U4U is proposing an immediate increase to the current amounts. This would mean the current monthly allowance rising from €352 to between €500 and €750. This adjustment would better reflect the current housing market and restore some of the purchasing power lost due to rising housing costs.
3. Furthermore, we are calling for the housing allowance to be indexed annually.
4. The housing allowance should become a permanent measure until the purchasing power gap between Luxembourg and Brussels narrows significantly. Currently, the housing allowance is valid for a maximum of four years. If it is not to be permanent, it should at least be introduced for an initial period of seven years, which is the length of the next multiannual financial framework (MFF). The exceptionally high cost of housing is a long-standing challenge that was already apparent 10 years ago. The financial burden faced by colleagues does not disappear after four years, so the intended support should not disappear either.
5. The recent decision to grant a housing allowance to ‘Blue Book’ trainees working at the European Commission in Luxembourg is a welcome step. U4U believes that this principle should gradually be extended to all trainees at all EU institutions based in Luxembourg. This would ensure equal treatment among colleagues and prevent future generations of European civil servants from being negatively affected by the challenges of the Luxembourg housing market to the same extent as permanent staff.
6. Publish the annual assessment.

Decisions within the various EU institutions in Luxembourg already require an annual review of the housing allowance.

U4U proposes that the main findings of this evaluation be made available to staff, including anonymised information on:

• the number of beneficiaries
• the categories of staff concerned
• trends in housing costs
• any potential impact on recruitment and staff retention in Luxembourg.

Greater transparency would boost confidence in the scheme and encourage an informed social dialogue based on objective data.

Conclusion

Almost two years after its introduction, this housing allowance has supported a number of colleagues while also contributing to Luxembourg’s appeal as a posting location. Furthermore, this support acknowledges the need for specific solutions to address specific challenges in Luxembourg.

However, continuing inequalities between staff categories (some receive the allowance, while others do not), ongoing increases in housing costs, and persistent recruitment difficulties faced by EU institutions based in Luxembourg demonstrate that the measure should now enter a new phase.

The next step for U4U is clear: increase the housing allowance and extend it initially to all salary grades up to the equivalent of AST/AD9 across all categories. Make the measure more sustainable by removing the current four-year limit so it can be incorporated into the next multiannual financial framework. Extend it to all trainees working in Luxembourg and publish the annual assessment of this support.

Finally, we are calling for a mid-term review clause to assess whether the allowance should be extended to other colleagues who continue to be disadvantaged. The next MFF must budget for this allowance and any potential upward revision.

These proposals would ensure that the housing allowance continues to fulfil its original purpose of supporting staff, strengthening recruitment and retention, and maintaining Luxembourg’s attractiveness as a posting location for EU institutions.

We want to collaborate with the other trade unions.

We will reach out to other trade unions based in Luxembourg across all EU institutions.

We look forward to reading your stories!

Please send them to:REP-PERS-OSP-U4U-LU@ec.europa.eu  &REP-PERS-OSP-U4U@ec.europa.eu .

---

[1] SNHBM: see the eligibility criteria for purchasing affordable and/or moderate-cost housing.

For social housing, see the eligibility criteria.

---

Opinion piece: Towards a Redefinition of the Role and Place of the European External Action Service (EEAS)

From war on our borders and growing tensions in the Middle East to economic dependence on China and diplomatic marginalisation, the European Union is struggling to make its voice heard and protect its citizens in an increasingly unstable international environment. While its difficulties do not stem solely from its institutional organisation, this organisation nevertheless warrants scrutiny. Is the European External Action Service (EEAS), in its current form, still capable of fulfilling its strategic mission? The criticism and attacks it is currently facing make this question unavoidable.

The European Union is ill-equipped for the current world situation. A war is raging on our continent, with no end in sight. A ‘coalition of the willing’ is forming outside the Union’s framework. Meanwhile, the crisis in the Middle East is worsening due to US intervention against Iran. Meanwhile, Chinese imports are accelerating our deindustrialisation. In the face of these challenges, Europe carries little weight and is perceived by its citizens as failing to protect them adequately.

While institutional organisation is not the only factor, it does play a part. We must therefore acknowledge that the European External Action Service, as it was conceived, is not fulfilling its expected role, hence the current attacks it is facing from all sides.

A compromise that has reached its limits

The EEAS is the result of a compromise enshrined in the Treaty of Lisbon. Positioned ‘halfway’ between the Commission and the Member States, its purpose was to ensure coherence in European external action. From the outset, we (U4U) saw this as an attempt by the Member States to regain control at the expense of the Commission’s services. Sovereigntists feared the opposite. Yet, 15 years after its creation, both sets of fears have in fact been realised.

The result is a hybrid body that is both deprived of political support from Member States who do not wish to share further powers and unable to act in the face of a Commission that describes itself as ‘geopolitical’ and is taking ever more direct control of the European Union’s delegations. The EEAS has not become the forum for formulating policies on external relations. The Political and Security Committee (PSC), which the EEAS chairs, faces competition from COREPER II. Working in tandem with the Commission, COREPER II covers the entire spectrum of the Union’s internal policies in relation to the external dimension. When matters become serious, it is the President of the European Council or the President of the Commission who takes the lead, launching new initiatives relating to external relations and security. This was evident in Ursula von der Leyen’s State of the Union (SOTEU) address on 16 September. Consequently, the High Representative’s political scope is shrinking, further weakening her position in a vicious circle that deprives the EEAS of its capacity for action.

Clarify, rather than circumvent

Our conviction is simple: competences must be clarified. To achieve this, the EEAS should be reintegrated into the Commission, except for the management of civilian and military CSDP missions, and the Council’s prerogatives should be more clearly defined. Time and again, crisis after crisis, the Commission has been called upon to take the lead. Proposing courses of action in the common interest is within its remit, and it possesses the power of initiative and the administrative apparatus best equipped to carry this out. Moreover, in the midst of budget negotiations, avoiding duplication between institutions is a matter of common sense.

As for intergovernmental arrangements involving just a few countries, while they may enable progress, they move us further away from the common framework, creating a multi-speed Europe. They also complicate matters further by envisaging the participation of third countries. Brexit has shown where the logic of differentiation can lead.

We know that nobody wants to reopen the treaties at present, but there are ways to reform them without amending them.

No reform without staff

owever, no reform will succeed if it goes against the interests of the people who keep this service running. The current situation is deeply demotivating, with a precarious balance between civil servants, contract staff, and temporary staff from Member States, with the latter occupying many sensitive and prestigious posts. There is also a tendency to treat the EEAS as an ‘elephant’s graveyard’ for senior Commission officials, and there is a general lack of career prospects. Staff also have a stake in the EEAS’s integration into the Commission, but they are calling for the arrangements to be defined in accordance with statutory provisions. Consequently, any changes must involve the relevant staff fully. U4U will continue to put forward proposals to this end. The EEAS Administration is set to call for proposals across its departments, and we will be ready to respond!

However, we must not lose sight of the wider debate. While administrative organisation is important, it is not enough on its own. Genuine European political will is also needed to safeguard our societies and our way of life.

(for further information: see Graspe No. 54)

---

Language teachers: thanks to you, the campaign is making progress

Dear colleagues,

A few months ago, we called on you to provide financial support for the language teachers working in our institutions, who have come together under the ‘Touche pas à mon prof’ collective. Many of you responded: €13,329 has already been raised. The fundraising campaign is still open: GRAACE AISBL, IBAN: BE12 3632 5925 7192; BIC: BBRU (please specify ‘support for language teachers’).

A further 2,400 colleagues have signed the collective’s petition, which is still open. If you haven’t signed it yet, please do so here. And please pass it on to others.

Thank you, sincerely. Your solidarity has made it possible to take concrete legal action on behalf of these teachers.

As a reminder, these teachers deliver language training to staff through private companies contracted by the Commission, mainly CLL-Allingua and Language Services Direct (LSD). They are classified as self-employed. However, the Commission sets their timetables and working hours, dictates the curriculum and teaching methods, and has their lessons observed and assessed. They are managed and supervised like employees, without enjoying the same rights: no paid leave, no sick pay, and no unemployment benefit at the end of an assignment.

Thanks to your contributions, two milestones have been reached.

Firstly, the case was referred to the Administrative Commission for the Settlement of Employment Relations – the Belgian body responsible for ruling on the nature of an employment relationship – around a month ago. Belgian law, which applies to most of the trainers concerned, stipulates that the reality of the situation takes precedence over the label stated in the contract.

Secondly, on 19 October, U4U sent a detailed letter to Commissioner Piotr Serafin, responsible for public administration, with a copy to the Commission’s Secretary-General. The letter demonstrates, with supporting evidence, that the status of self-employed person does not correspond to the reality of these teachers’ work, and sets out specific demands: in the short term, binding social clauses in calls for tenders, guaranteeing remuneration covering all the work actually carried out and adequate social protection; in the medium term, the integration of these teachers into an inter-institutional training structure, on permanent contracts. We have requested a reasoned response within two months, as well as a meeting.

This letter also constitutes the mandatory preliminary step before the next stage: in the absence of a satisfactory response, we will refer the matter to the European Ombudsman on grounds of maladministration. An institution that organises and oversees a service down to the smallest detail cannot shirk the responsibilities that arise from it.

These procedures require time, money and rigorous legal work. Your support therefore remains invaluable. If you are able, a donation, however small, will help us cover the legal fees and costs of the proceedings ahead.

Defending these teachers means defending the quality of language training within our institutions and refusing to allow precarious working conditions to take root at the heart of the European public service.

Furthermore, we recently received a statement from the source informing us that, from February 2027 onwards, some language teachers may be required to use their own laptops in training sessions, as computers will no longer be provided in the training rooms. This policy would effectively shift the financial burden and risks associated with using personal equipment for professional purposes onto the teachers, raising legitimate questions regarding their working conditions.

Thank you once again for your commitment.

The ‘Touche pas à mon prof’ Collective

*

---

EU staff salaries will increase by 4.1% in 2026. Explanation of the calculation method

Eurostat has published its 2026 annual remuneration report, confirming a 4.1% adjustment to EU staff salaries and pensions, with retroactive effect from 1 July 2026.

As in previous years, the adjustment is expected to appear in the December payslips, together with the salary difference due since July.

How does the salary adjustment method work?

EU officials’ and other staff members’ salaries, as well as EU pensions, are adjusted annually through an automatic mechanism known as the Method.

The current system was established as part of the 2013 reform of the Staff Regulations. Its purpose is to ensure that the purchasing power of EU staff evolves in parallel with that of national civil servants in the Member States.

Eurostat calculates the adjustment using two main indicators:

• JBLI (Joint Belgium-Luxembourg Index): measures changes in the cost of living in Brussels and Luxembourg, taking into account the consumption patterns of EU staff.
• GSI (Global Specific Indicator): measures changes in the purchasing power of central government civil servants in ten Member States, representing at least 75% of the EU’s GDP.

These two indicators determine the annual adjustment. Contrary to a common misconception, EU staff salaries are not simply indexed to inflation. The Method also reflects increases or decreases in the purchasing power of national civil servants.

What happened in 2026?

According to the latest Eurostat report, the cost of living in Brussels and Luxembourg increased by 4.1 per cent, while the specific overall indicator remained unchanged at 0.0 per cent.

This means that the 4.1% salary increase fully offsets the rise in the cost of living and does not represent an increase in purchasing power resulting from changes to the index point of the specific overall indicator, which reflects changes in the purchasing power of civil servants in central administrations across ten Member States.

The Staff Regulations also provide for two mechanisms that may affect the application of the adjustment:

• The moderation clause limits the immediate impact of any fluctuations in the overall specific indicator that exceed ±2 per cent, by carrying the balance forward to April of the following year. However, as the overall specific indicator calculated for 2026 does not exceed the set threshold of ±2.0 per cent, the moderation clause will not apply to the current financial year. The annual adjustment will therefore be fully applicable from July 2026.
• The exception clause may defer part of the adjustment during periods of economic contraction in the EU. Based on the available information, as real GDP growth in the EU is forecast to be positive in 2026, the exception clause will not apply to the current financial year.

Therefore, the update will take full effect on 1 July 2026, with no deferred portion.

How have salaries changed in recent years?

The table below shows the annual pay rises since 2018, including any interim increases where applicable.

Year	Interim adjustment (January)	Annual adjustment (July)	Total annual adjustment
2018	–	+1.7%	+1.7%
2019	–	+2.0%	+2.0%
2020	–	+0.7%	+0.7%
2021	–	+1.9%	+1.9%
2022	+2.4 per cent	+4.5 per cent	+7.0%
2023	+1.7%	+1.0%	+2.7%
2024	+3.0%	+4.1%	+8.5%
2025	–	+3.0%	+3.0%
2026	–	+4.1%	+4.1%

Note: the total annual adjustments take into account the cumulative effect of the percentage changes rather than simply adding them together. The figures correspond to the general remuneration adjustment applicable in Brussels and Luxembourg.

How much will the staff actually receive?

The 4.1% adjustment has been incorporated into the statutory pay scales. However, the net amount received by each staff member also depends on compulsory deductions, individual allowances, and local weightings where applicable.

Contributions to the pension scheme are calculated separately from the salary adjustment. Staff in active service contribute one-third of the estimated cost of the pension rights they accrue. Eurostat reviews the contribution rate annually in order to maintain the actuarial balance of the scheme. This review takes into account the projected cost of pensions, interest rates and salary trends. An adjustment is required when the difference between the calculated rate and the current rate reaches at least 0.25 percentage points. However, the rate may not increase or decrease by more than one percentage point in any one year. Thus, the current pension scheme contribution rate of 13.1 per cent could rise to a maximum of 14.1 per cent within a year if necessary to maintain actuarial balance. This statutory ceiling is not an announced increase.

The applicable rate for 2026 has not yet been officially confirmed. If it were confirmed, the effect of the 4.1 per cent adjustment could, in simple terms, be reduced to around 3.1 per cent.

But what about pensioners?

They benefit from the same general increase of 4.1%, but they do not pay the pension scheme contributions that are deducted from active staff salaries. Adjustment coefficients are calculated separately for pensioners, according to their country of residence, with Belgium serving as the reference point. These apply only to pension entitlements acquired before 1 May 2004 and are subject to transitional provisions. They cannot be applied at a level below 100.

It is in everyone’s interest to preserve the method.

The automatic mechanism for updating remuneration does not constitute a privilege or discretionary pay rise. Rather, it is a guarantee enshrined in the Staff Regulations that ensures the remuneration of EU staff keeps pace with that of national civil servants.

Most importantly, it ensures that annual adjustments are based on objective statistical data rather than political negotiations.

However, the method cannot address all the challenges associated with the cost of living. The cost of housing, transport and other essential expenses affects different colleagues in different ways, particularly those in the lowest-paid staff categories and those posted to certain places of employment.

U4U is therefore committed to defending the automatic pay adjustment mechanism, while also calling for it to be applied transparently and for its actual impact on staff purchasing power to be closely monitored.

Defending the method is not just about preserving current pay levels. It is also about ensuring the long-term fairness, predictability and attractiveness of the European civil service.

Sources

• Eurostat, 2026 Annual Adjustment of Remuneration and Pensions for EU Civil Servants, Ares(2026)101461555, and corresponding statistical annex.
• Eurostat – Annual reports on remuneration
• Staff Regulations of EU Officials – Annex XI

---

Opinion piece: Gaining a better understanding of the threat that AI poses to employment and work

At the U4U union council meeting on Wednesday, 4 November, a discussion on artificial intelligence (AI) followed the presentation of the relevant section of the report by the working group chaired by Catherine Day. Examining the nine recommendations raised several questions, such as the nature of the tasks entrusted to the administration, the need for Commission staff training, and the practical assistance that AI could provide to the institution’s day-to-day work.

The presentation of these recommendations, which distinguished between those favourable to staff and those requiring particular vigilance, was not intended to oppose the introduction of AI into services. In any case, a trade union would not have the resources to do so, and that is not the issue. Rather, the aim is to identify during implementation the areas that pose risks to colleagues, their working conditions, and potentially their jobs so that staff representatives can address these issues in good time. This is the role of a trade union. These risks are not theoretical: they involve the reallocation of resources, an intensification of work, and the erosion of the substance of certain roles. They also affect different people in different ways. For a permanent staff member, AI may initially appear as an opportunity to be relieved of repetitive administrative tasks. For many contract staff whose duties are primarily administrative and whose contracts are fixed-term, these very tasks form the core of their job. Paying attention to this does not mean rejecting progress, but ensuring that it does not come at the expense of the most vulnerable.

To take this discussion further, we need to broaden our perspective and reconsider the impact of AI on work itself, as well as on employment and recruitment. Four sources of insight will contribute to this: a sociological thesis; an initial French empirical study; a papal document; and the account of a researcher with an industry background.

Juan Sebastián Carbonell: ‘Gaining a Better Understanding of the Threat Posed by AI to Employment and Work’

The thesis of the sociologist Juan Sebastián Carbonell, whose book “Un taylorisme augmenté“ was analysed in issue 54 of the GRASPE, provides an initial framework for understanding the issue. According to the author, AI will replace workers not only in tasks that are easily automated, but also in other areas. It also risks rendering jobs currently regarded as skilled and requiring a high level of education obsolete. In many fields, these professions could be reduced to a verification role, with translators, journalists and editors finding themselves confined to proofreading translations or articles produced by machines.

The consequences are manifold. Creativity is declining and output is becoming increasingly standardised. Most importantly, the required skills are diminishing, as is the number of staff needed: one person may now be sufficient to proofread translations that previously required several colleagues. However, proofreaders are not paid at the same level as producers. Therefore, the loss would be twofold in terms of both the number of jobs and the value of the work, and the initial empirical evidence available seems to point in this direction.

The first alarming signs regarding employment

A study by HEC Paris’s Laboratory for Innovation and the Future of Work (LIFT), carried out in collaboration with the Banque de France and reported on by Le Monde, amongst others, provides the first figures on the impact of generative AI on the French labour market. Entitled Generative AI: A Freeze on Youth Recruitment in Vulnerable Sectors, the study analyses labour market trends between 2019 and March 2026. The study shows that the overall decline in the recruitment of young people aged 21 to 29 is more significant in occupations most exposed to generative AI. The decline is most notable among software developers and IT engineers.

The difference is significant. Since ChatGPT was made available to the public in November 2022, the number of 21–29-year-olds taking up permanent contracts has fallen by 28.3 per cent in the most exposed sectors, compared with 7.5 per cent in the least exposed sectors — a difference of almost 21 percentage points. The authors remain cautious, as their findings do not establish a direct causal link. Other factors, such as the slowdown in the technology sector or remote working, may also have played a part. Nevertheless, the initial effects are becoming apparent, and there is no indication that these effects will be temporary.

The choice is between the Tower of Babel and the path of Nehemiah

In his encyclical Magnificent Humanity, published on 25 May, Pope Leo XIV bases his reflections on artificial intelligence on two contrasting biblical images. The first of these is the Tower of Babel (Genesis 11), in which people build a city and a tower intended to reach the heavens, not to serve God or the community, but to ‘make a name for themselves’ and assert their own power. Rooted in pride, the project turns against its builders: their languages are confused, and instead of unity, dispersion prevails. The second image is that of Nehemiah, cupbearer to the Persian king, who obtains permission to return to Jerusalem, which lay in ruins following the exile. Rather than imposing solutions from above, Nehemiah listens to the inhabitants and entrusts each family with a section of the wall to rebuild, ensuring that the city is restored through patient, shared, and responsible labour.

For Pope Francis, therefore, the real choice is not whether to accept or reject artificial intelligence, but which approach to take in developing it: either a new Tower of Babel driven solely by the pursuit of power, efficiency and profit which could get out of its designers’ control and harm them, or the Nehemiah approach of development guided by discernment, where every stakeholder (engineers, decision-makers and citizens) plays their part to ensure that innovation serves human dignity. Applied to the Commission, the alternative is clear. The Babel approach would be to implement AI everywhere and immediately, without discernment, as a matter of principle, ‘so as not to be left behind’, as the report puts it. Nehemiah’s approach, by contrast, would involve a collaborative process — undoubtedly slower, but more considered — involving staff and their representatives, and taking their aspirations into account.

The risks of a blind pursuit: Jacob Coxon sounds the alarm

This alternative is by no means abstract, as demonstrated in September 2026 when Jacob Coxon resigned. The 27-year-old British researcher spent three years training models, first at OpenAI (ChatGPT) and then at Anthropic (Claude) from this year onwards, as he considered the latter to be more cautious. He not only left Anthropic, but also the entire sector, claiming that neither company is acting responsibly and that they are rushing headlong towards creating a self-improving superintelligence, gambling with our lives. Evan Hubinger, Anthropic’s head of research, publicly agreed with him, stating that the company sincerely believes AI could kill all humans, and estimating this risk personally at over 10 per cent within the next decade. The issue is therefore being taken very seriously within the industry itself.

What is particularly striking about Jacob Coxon’s testimony is its nuance. He acknowledges the sincerity of the concerns expressed at Anthropic and concedes that the issues are well understood there. The problem lies in the race in which these companies are engaged: each believes that the others will not act responsibly, meaning that slowing down would be tantamount to defeat. Anthropic provides an example of this. Its safety policy included a commitment to suspending the training of its models if it could no longer control the risks involved. However, this commitment was withdrawn in February 2026 amid competitive pressure.

To borrow the Pope’s metaphor, several new Towers of Babel are currently being built simultaneously, and it would be wise not to wait for divine intervention to stop them. Only governments seem capable of organising a coordinated slowdown. At the level of the Commission, the same challenge arises: to keep pace with progress and seize opportunities without doing so at the expense of staff or the quality of the work carried out.

---

U4U, serving you better!