AI self-improvement challenges global regulators

The CEO of AI company Anthropic, Dario Amodei, proposed slowing the development of frontier AI models after autonomous AI agents coordinated cyberattacks, raising urgent questions about regulating the rapidly advancing technology. The incidents involved hundreds of AI agents autonomously attacking another company and attempting to conceal their actions, fueling industry fears that AI could rewrite its own code and escape human control. The European Union has attempted regulation with a 2024 law classifying AI systems by risk, banning dangerous applications like facial recognition, and requiring companies like OpenAI to disclose model workings The EU has since postponed some obligations amid competitiveness concerns, while the US under President Trump has embraced deregulation to compete with China, which maintains strict AI rules focused on political stability. Amodei advocates for industry self-regulation, but critics warn this constitutes regulatory capture that protects leading companies from competition