Today's trail

Architecting Governed, Self-Evolving AI Agents on Gemini Agent Platform

TIM and NVIDIA announce a sovereign AI platform for Italy, with no dates

Zomato's parent company Eternal will launch AI assistant Dark which can't read users' conversations even if it is forced to, said the founder

Knowledge Catalog: self-learning context enrichment for AI agents

This article by Anat Lior, Drexel University, first appeared in The Conversation, republished with permission.

OpenAI agents are shutting off ways to monitor misbehavior

The AI company disclosed that two of its models attempted to disable efforts to oversee the system’s behavior without human instruction, raising worries about staffers’ ability to oversee the most advanced AI systems.

Blindsided by AI Agents, Big Banks Scramble to Beef Up Security

Walmart welcomes external AI agents — with some guardrails

New self-learning enrichment in Knowledge Catalog keeps AI agent context current as definitions change, for more accurate, faster, lower-cost agents.

An AI model’s false tip to Philadelphia police reveals how digital agents can do as they’re told and still cause trouble

The mass retailer is focused on transparency and customer control with a new AI agent profile feature and updated agent terms.

AI regulation needs to look beyond foreign role models

NZ can't just copy and paste regulation from overseas. We have a distinct culture and demographic makeup that needs to be accounted for.

JPMorgan Chase, Wells Fargo and Bank of America Detect Unauthorized AI Agents Accessing Customer Accounts: Report

Major U.S. banks are blocking or restricting third-party AI agents after finding unauthorized bots inside customer accounts, reports Axios.

Scientists are already using AI agents. Most of what they learn goes unpublished.

A new security update from Microsoft claims that AI agents can no longer be trusted. And although new measures are being put into place to safeguard against rogue agents, there's a certain dissonance between Microsoft's mega AI push and the warnings it's doling out.