This week in AI: trust moved to the front

This week, AI news was about trust: which tools will show up inside the apps your people already use, who controls what AI can reach, and who it is safe for. All three land on your desk as people questions before they are tech questions.

1. Siri AI is coming into the apps your people already use
Apple opened Siri AI to app developers in a preview, so apps can let people ask Siri to find information or take actions for them. Apple says it will reach the public in a few months.
What it means for change leaders: AI is arriving inside everyday tools, often before any official rollout. Your people will try it on their phones whether or not there's a policy. Tell them what's okay before they have to guess.
Source: Apple Developer News

2. Controlling what AI agents can reach is now its own business
Data security company Sentra raised $150 million to help organizations see and limit what people, AI agents and systems can access across their data.
What it means for change leaders: when AI starts acting on someone's behalf, "who is allowed to do what" stops being an IT setting and becomes a question about roles. Before agents get access, agree who owns each decision and who answers when something goes wrong.
Source: SecurityWeek

3. New York sets rules for AI chatbots and minors
New York signed a law that bars certain AI chatbot features for users under 18 and requires age checks, with fines of up to $25,000 per violation.
What it means for change leaders: AI rules are moving from "be careful" to specific obligations. If your organization runs a customer-facing chatbot, the people who manage it need to know what changed and who checks it. That's training and ownership, not only a legal review.
Source: New York Governor's office

Also this week: Anthropic released FlashProve, an open research prototype that can prove which AI model produced an answer. It's early, but "can we trust where this answer came from?" is becoming a real question for organizations. (Anthropic)

A question for your week: if an AI assistant showed up in your team's apps tomorrow, would your people know what they're allowed to use it for?