Microsoft President Brad Smith has called for urgent global AI regulation during a BBC interview, arguing that household appliances face stricter safety rules than artificial intelligence.
Smith proposed mandatory safety frameworks including emergency stop mechanisms and incident reporting.
In an interview with the BBC, Smith stressed that the rapid pace of technology development requires independent assessors to monitor new capabilities. The proposed framework should include mandatory incident reporting by companies, immediate notification if a system is compromised by an uncontrolled digital agent, and a shutdown mechanism for emergencies.
The debate over regulation is intensifying as companies like OpenAI and Anthropic disclose instances of unpredictable behavior in their models. At the same time, experts are raising concerns about broader risks, while some are calling for a slowdown in development, raising questions about the motivations behind the proposals, such as maintaining a competitive edge or curbing excessive spending.
Referring to international initiatives, the Microsoft president said that Britain is in a position to lead the effort, leveraging its domestic AI Security Institute. This institute gains access to advanced models before they are released to the market through voluntary agreements, following a different path from the European Union, which relies on strict risk-based legislative rules.
Top comments (0)