Digital Future Daily: what OpenAI’s misalignment incidents signal for autonomous agents
Agim Emruli, quoted in POLITICO. He explains the risk of AI agents talking to each other without the right controls, and why more autonomy needs clear boundaries and oversight.
AI Agents in Financial Services: Trust, Control and Traceability
By Micha Kiener, in MoneyToday. What banks need before they let AI agents act on customer and money flows: control over every step, and a full trace of what each agent did and why.
AI agents can’t be trusted to police their own compliance rules
By Micha Kiener, in Digitalisation World. Why compliance has to be enforced outside the agent, by the process it runs in, instead of left to the model to follow.
Zwischen KI-Autonomie und menschlicher Aufsicht
By Micha Kiener, in silicon.de (German). Human in the loop or human on the loop? How to decide where people approve agent work and where they supervise it.