Saturday · 2 stories

10 October 2026

Red emergency stop button on a metal pole — Photo: Tobias Westphal on Unsplash

Microsoft's CEO says advanced AI needs an emergency brake that humans control

Design principles listed: 6
Posted: Sat 10 Oct; on X

Microsoft CEO Satya Nadella said on 10 October that advanced AI systems should be built with containment, independent controls and an “emergency brake” that lets authorised people pause or shut down a model mid-task, CNBC reported.

In a post on X, Nadella listed principles including model diversity, a human-readable record of what a model does, continuous testing, auditability, containment and incident disclosure. He said frontier models, closed and open, should be treated like insider risks.

His comments follow warnings about AI safety from other tech leaders, including Anthropic's Dario Amodei and OpenAI's Sam Altman, CNBC said.

Person using a laptop while holding a card — Photo: rupixen on Unsplash

Musk tells users to photograph their credit card so Grok can shop for them

Visa: bot-initiated fraud: +25%; 6 months to Nov 2025
In the US: +40%

Elon Musk wrote on X on 10 October that users could take a picture of their credit card, drop the photo into a chat and let Grok find the best deal on an item and order it. He later posted an order of Lego Star Wars kits that he said Grok had placed.

Community Notes under the post said xAI's own policy tells users not to put personal information in prompts, Moneywise reported. Moneywise said it was unclear whether card-photo shopping is a new feature or which Grok tier allows it.

Visa said in November 2025 that fake storefronts can trick AI shopping agents into handing over stored card details. Visa reported a 25% rise in malicious bot-initiated transactions over six months, and 40% in the US.

All dates