Can AI Companies Really Police Themselves?

AI companies are promising to police themselves — but can voluntary safeguards really protect society from the risks of increasingly powerful artificial intelligence?

Shelly Palmer, Professor of Advanced Media in Residence at Syracuse University’s Newhouse School of Public Communications, joined CNN International live on Wednesday, September 30, 2026, to discuss President Donald Trump’s White House meeting with leading AI executives and the new voluntary agreement calling for greater self-policing of frontier AI development.

Palmer pushes back on sensationalized descriptions of AI models “going rogue,” explaining the important distinction between frontier AI models, experimental AI agents, open-weight and open-source systems, and tools deliberately deployed by bad actors. He argues that the companies building the most advanced commercial AI models are uniquely positioned to control how quickly those systems develop and what safeguards are built around them — while malicious uses of freely available AI present a very different challenge.

The conversation also addresses the larger, harder-to-answer risks of artificial intelligence: what increasingly capable AI systems could mean for society, the economy and human decision-making. Palmer argues that these uncertainties deserve serious attention, while warning that confusing unusual experimental agent behavior with autonomous “rogue AI” can obscure the real issues.

Can the AI industry effectively regulate itself? What happens when powerful AI tools fall into the hands of bad actors? And are we focusing on the right risks as AI capabilities accelerate?

Original Airdate: September 30, 2026

Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.