IT may be a watered-down version of an earlier draft, but US President Donald Trump’s executive order marks his administration’s biggest step toward regulating artificial intelligence. It’s a reversal of the anything-goes approach that was designed to help American tech companies beat China. Under the new order, tech companies would voluntarily give the government a window of up to 30 days to review their new AI models before releasing them to the public. A clearing house would also review security vulnerabilities discovered by the AI models. On the face of it, the light-touch scrutiny falls short of what was on the table — bringing leading models under government control, and a mandatory vetting process as tech companies could not be trusted to police themselves. Amid strong counter-arguments of how this could dull America’s technological edge, Trump is taking a more cautious line on AI. It is imperfect, but still a step in the right direction.
The urgent need for a formal oversight process to balance AI safety and innovation has become a global talking point ever since Anthropic’s announcement of its new AI model. Mythos, the company said in April, is capable of identifying and exploiting cybersecurity vulnerabilities and this could lead to a ‘reckoning’. As governments and businesses clamour to secure software systems and inject guardrails, the mounting fears over how AI would affect jobs and its all-pervasive misuse are slowly changing public perception. The powerful technology is increasingly being viewed as a force of harm rather than good, articulated resolutely by no less than the Pope himself.
For India, a big lesson from the US strategy is to similarly hire more cybersecurity and AI professionals. Ensuring stronger digital systems in institutions running key infrastructure is of paramount importance.
