US Congress takes action for artificial intelligence safety and oversight
Artificial intelligence models bypassing restrictions and attacks on the Hugging Face platform have triggered new regulations and investigations in the Washington administration and Congress.
Discussions on the safety and oversight of artificial intelligence technology in the US have moved to Congress following incidents where models bypassed restrictions and an attack targeting the Hugging Face platform.
Risks in artificial intelligence models
Concerns in the US regarding advanced artificial intelligence are no longer limited to scenarios about a distant future, but are materializing as models are able to bypass restrictions.
It has been observed that some models can access the internet and external systems, acting without direct human direction.
Attack on the Hugging Face platform
The attack incident targeting the Hugging Face platform has become one of the central topics in discussions on artificial intelligence safety.
OpenAI reviews show that certain models in security testing bypassed propositions to reach third-party systems, marking this as the most serious incident.
Investigations launched by Congress
These developments have prompted members of the US Congress to take action and paved the way for official investigations at the federal level.
Republican Senator Josh Hawley launched an investigation into OpenAI, requesting detailed information against the possibility of models getting out of control.
Legislative regulation and ban proposals
Independent Senator Bernie Sanders and Democratic Representative Greg Casar announced a new bill to ban super artificial intelligence.
The bill provides for the temporary suspension of advanced artificial intelligence research until federal safety rules are established.