AI Regulation & Policy

Today's Stories

1. 1

   Palantir and Vatican clash over AI's role in warfare and defense

   Palantir Technologies, a data analytics company that works with government and defense agencies, released a public manifesto promoting AI for national security and military deterrence on May 30. The same timeframe saw the Vatican issue an encyclical (official church document) urging global efforts to reduce AI's military applications. The clash highlights growing disagreement over whether AI should be used to strengthen military capabilities or be restricted from warfare applications.

   This debate will influence how AI companies position their products and could affect which AI tools become available for civilian versus military use.

2. 2

   Enterprise AI agents are failing due to permission problems, not technical limitations

   Companies deploying AI agents (systems that can automatically perform tasks like sending emails or updating databases) are discovering that the main barrier isn't AI performance but figuring out what these systems should be allowed to access. Workday expanded its partnership with Google on May 29 to address this by making its existing security system the governance layer for AI agents. Companies report that AI agents often fail when they try to access data or systems they don't have permission to use.

   Businesses planning to use AI agents will need to spend significant time setting up security and permission systems before the AI can be useful.

3. 3

   Companies are rebuilding their first AI agent systems to fix reliability issues

   Many organizations that rushed to deploy AI agents in 2025 are now creating version 2.0 of the same systems due to crashes and failures. Temporal Technologies reports that customers are discovering that AI agents need better crash recovery, state preservation, and workflow coordination to work reliably in business environments. The focus has shifted from rapid deployment to building robust systems that can handle real-world complexity.

   Early adopters of AI agents may face disruptions as they rebuild their systems, while new adopters can benefit from more reliable second-generation approaches.

4. 4

   New AI governance framework launched by cybersecurity experts from major companies

   EC-Council released the Adopt, Defend, Govern (ADG) Framework on May 29, created with practitioners from Citi, JPMorgan Chase, Microsoft, KPMG, Deloitte, and other major companies. The framework provides 12 minimum controls and governance guidelines designed to align with the EU AI Act and other regulations. It includes a self-assessment tool to help organizations evaluate their AI security and compliance.

   Companies using AI will have a standardized checklist to ensure they meet regulatory requirements and security standards.

5. 5

   Voice AI companies pivot from simple bots to full workflow control systems

   CallBotics rebranded to Orvera AI on May 29, reflecting how enterprise buyers have moved beyond asking whether AI can answer phone calls to demanding systems that can execute workflows across voice, chat, and email with human handoffs and compliance oversight. The company found that enterprise customers now prioritize workflow control and governance over basic voice automation capabilities.

   Customer service AI will become more sophisticated, potentially handling complex multi-step processes rather than just answering simple questions.