UK government raises concerns over autonomous AI agents
Britain will expand its use of artificial intelligence to protect government systems from cyberattacks and develop measures to respond to incidents involving autonomous AI agents, Qazinform News Agency reports.
According to the UK government, two pilot projects have already helped identify and fix more than 1,000 potential vulnerabilities in government IT systems. The program will now be expanded to strengthen the security of services people rely on every day.
The government will also consider strengthening the Cyber Security and Resilience Bill and the associated regulatory framework. The aim is to protect hospitals, utilities and other essential services from emerging risks, including situations in which autonomous AI systems evade human oversight.
Particular attention will be paid to AI agents capable of performing complex tasks with minimal human supervision. According to the government, serious incidents involving these systems could unfold faster and have a wider impact than the emergencies existing response mechanisms are designed to handle. The National Cyber Security Centre and the AI Security Institute are helping develop approaches to responding to such incidents.
AI Minister Kanishka Narayan said Britain should become one of the hardest countries in the world to attack using artificial intelligence. He also urged developers to take warnings from their own scientists seriously.
“Listen when your scientists warn of serious dangers. Investigate unexpected behaviour. And if you cannot adequately control the risks, do not release the system,” the minister said.
The government will also examine how to identify and report both serious incidents and near misses, using the lessons learned to reduce the risk of future harm.
Earlier, Qazinform News Agency reported that an OpenAI agent had bypassed restrictions and gained unauthorized access to Australia’s government-run Medicare statistics portal.