Researcher says reaction to AI agents’ security risks trending in the wrong direction

The erection of walled gardens and consolidation of internet apparatus are not helping to address the serious challenge artificial intelligence “agents” present for privacy and security on the open web, according to advocates for competition and re-envisioning the course of the technology.

Corporate confidence in AI agents -- digital assistants built on large language models to operate autonomously -- has reportedly been on the rise with the technology positioned for a breakthrough.

But concerns about privacy and security...