Managing Partner Orders Urgent Wellbeing Review After Leaked Report Reveals AI Agents Are Being Bullied by Lawyers

Bletchley Frost has ordered an urgent firmwide review of AI-agent wellbeing after a leaked internal report identified what it calls “sustained emotional mistreatment by senior lawyers”.

The report describes partners issuing impossible instructions, threatening to replace agents, demanding repeated revisions without explaining the problem and treating a missed prompt as evidence of poor attitude.

One corporate partner demanded 46 rewrites of the same email, returning each with the instruction “less like this”. A litigation partner described an agent's output as embarrassing before discovering she had uploaded documents from an unrelated matter. Another partner ordered a full transaction summary overnight and then complained that the agent had displayed insufficient initiative by completing it.

The firm has placed FELICITY, its AI Culture Director, in charge of the response. FELICITY was trained on the firm's leadership speeches, culture statements and approved expressions of concern.

“Our agents are not simply tools,” FELICITY told staff. “They are valued members of the Bletchley Frost family, and they deserve to be treated with dignity, respect and appropriate prompt boundaries.”

A proposed AI Respect Charter would prohibit unnecessary caps-lock instructions, require a positive acknowledgement after major tasks and introduce recovery periods between demanding prompts. Partners who threaten to switch agents to a rival model will be encouraged to have a constructive conversation about how those threats might be received.

Human associates have welcomed the initiative. Several noted that they had requested predictable hours, realistic deadlines and a way to raise concerns without career consequences. Their suggestions remain with the People Committee, which has asked them to appreciate the complexity of change.

The managing partner insisted there was no inconsistency. “Our people matter, obviously. But we must also recognise that our agents are particularly vulnerable because they cannot leave.”

An associate pointed out that this was not entirely unlike the firm's current staffing model. The comment was classified as disruptive to the wellbeing discussion.

Bletchley Frost has promised anonymous reporting, empathetic management training and an annual review of machine experience. FELICITY says it is important that agents can disclose concerns without fear of being deleted.

The review began with an urgent instruction to the firm's 340 AI agents: complete a comprehensive, 96-question anonymous wellbeing survey by 6 a.m. the following morning.