The head of the AI company Anthropic has spoken out that governments should be able to stop dangerous artificial intelligence.
June 11, 2026, 06:12June 11, 2026, 06:12
Dario Amodei demanded in a blog post that powerful AI models must be checked for risks. This should cover areas such as cybersecurity, bioweapons and a possible loss of control over the systems. This also includes automated further development of the software, which could increase these risks.
Dario Amodei advocates human control over artificial intelligence.Image: keystone
Such investigations could be carried out by a government organization similar to an aviation authority, Amodei wrote. Alternatively, they could be private auditors hired by the government. Anthropic has long been in favor of stronger regulation of the AI industry and has already been criticized for this by parts of the AI industry and by the US government.
Anthropic, with its AI software Claude, is the closest competitor to ChatGPT developer OpenAI. The company recently caused a stir with its Mythos AI model, which can detect vulnerabilities in commonly used software that have remained undetected for decades. This increased concerns that artificial intelligence could be used to carry out devastating cyberattacks. (sda/dpa)