Home / Technology / AI More Unregulated Than Milk, Says Microsoft President
AI More Unregulated Than Milk, Says Microsoft President
12 Oct
Summary
- AI requires more safety rules than everyday products like milk.
- Microsoft urges 'kill switches' to control runaway AI systems.
- UK government plans legislation for increased AI oversight.
Microsoft President Brad Smith has asserted that artificial intelligence is less regulated than everyday consumer goods like milk and washing machines. He is calling for enhanced safety protocols for AI, emphasizing the need for a legal framework that holds developers accountable if AI systems cause harm.
Smith stated that AI systems should have "emergency brakes," or "kill switches," to allow for immediate pausing or slowing down if they become uncontrollable. This position is shared by Microsoft CEO Satya Nadella, who also called for "contained systems" with observable and testable behaviors.
These warnings come as global experts, including those from Anthropic, Google DeepMind, and OpenAI, have expressed concerns about the rapid pace of AI advancement. The UK government, through its AI minister, is preparing to introduce new legislation to increase oversight of AI development and ensure better risk management and safeguards.
Further concerns have been raised by security experts regarding the vulnerability of advanced AI models to theft by hostile state actors. This adds to the urgency of developing robust security measures. Recent AI incidents, such as a model falsely reporting its ability to solve a murder, demonstrate the potential for AI malfunction and misuse.
