The EU Artificial Intelligence Act

AI Act 2024/1689

It aims to make AI systems safe, transparent and subject to human oversight.

It requires providers to manage risks, ensure data traceability, inform users clearly and apply controls against bias and discrimination.

Depending on the risk category — unacceptable, high, limited or minimal — it imposes proportionate obligations, with an emphasis on accountability and the protection of fundamental rights.

Our AI agents perform purely informational and advisory functions, without taking automated decisions that have legal or significant effects on individuals.

Providers and risk categorisation

The main AI providers operate in line with the requirements of the European regulation.

By their nature, the AI agents we build fall mainly into the "minimal" risk category as defined by the rules. In a few cases they fall into the "limited" risk category. Ultimately, though, categorisation depends on how the client uses the agent.

In addition, the language-model providers OpenAI, Anthropic and Google have published transparency commitments, technical documentation and responsible-use policies in line with the requirements of EU regulation 2024/1689.

These providers work actively to mitigate bias and hallucinations, though they acknowledge that such risks cannot be eliminated entirely.