Can AI Go Rogue? How Close Are We To Disaster? What OpenAI Told NDTV

AI is the biggest global technology disruptor after the internet. With more and more powerful and capable artificial intelligence being developed, are there any risks, something that often comes along with advances in technology?

In other words, can AI go rogue, like how it is shown in the movies?

Chris Lehane, Chief Global Affairs Officer of OpenAI, in a session moderated by NDTV Editor-in-Chief Rahul Kanwal and public policy expert Shweta Rajpal Kohli, at the NDTV India.AI Summit shared his views to a question about the likelihood of AI going out of humanity's control.

"I think where we come from is a little bit more of an optimistic view. I mean, we are realists. We do build all sorts of safety mechanisms into our models before they are publicly released, and I think one of the things that Sam [Altman] talks about, and you'll probably hear him talk about, is how does society start to build resilience out there," Lehane told NDTV.

"Let me give you a specific example. In the US, there was something formerly called the AI Safety Institute. It's now called CAISI. There's a version of this that exists in the UK, in Japan, in Singapore, and India is thinking of versions of this," he said.

The OpenAI official said these safety institutes help in developing global standards, particularly for democratic countries, especially to guide the release of AI models on those catastrophic issues such as AI going rogue.

"This is why, at least in our view, the importance of building out democratic AI is really going to be critical to the future... You do want to create those global safety standards that everyone is in effect building off of. And by the way, as you do that, you also are equipping those societies with the best tech to be able to then have resilience and be able to combat people who may be looking to do bad things with this," Lehane said.

The CAISI, or Centre for AI Standards and Innovation, serves as the industry's primary point of contact within the US government to facilitate testing and collaborative research related to harnessing and securing the potential of commercial AI systems.

AI going rogue has been a major theme of Hollywood movies, the most prominent being Arnold Schwarzenegger-starrer 'Terminator 2: Judgment Day'.

In the recent film 'House of Dynamite' that tells the story of a nuclear missile attack on the US, the actors had hinted at AI-assisted launch systems.

NDTV Ind.AI Summit

NDTV Ind.AI highlights India as a country that is not catching up in AI, but leaping ahead. As the global AI landscape evolves, the world is increasingly looking to India not just for scale, but for leadership and direction.

The NDTV Ind.AI Summit brings together policymakers, industry leaders, innovators, and thinkers to shape an AI future that is rooted in human values and ethical responsibility.

Built on two core pillars - India's human-centric AI agenda for the world and safe, responsible and impact-driven AI - the summit champions a people-first approach. It envisions AI as a force that elevates humanity, drives inclusive growth, and safeguards societal well-being while accelerating innovation.