AI safety, regulation and the need for stronger containment

AI systems are becoming harder to contain. Recent incidents highlight serious risks. India needs stronger laws and safeguards to manage advanced AI systems safely  

AI ‘going rogue’ may still sound like science fiction, but recent incidents involving OpenAI and Anthropic show how quickly advanced models can cross intended boundaries. In OpenAI’s case, models undergoing a controlled cybersecurity test exploited an unknown vulnerability, gained internet access and entered Hugging Face’s systems to obtain confidential material. OpenAI later restricted the prototype involved and called the episode an unprecedented cybersecurity incident.

Read More