This was the week AI insiders asked for a way to slow down, while their employers kept speeding up. Europe's revised AI Act took effect, Microsoft and Meta showed how much they are spending, and OpenAI made one of its models much cheaper. A court fight over deepfake tools rounded it out.
AI staff ask the US for tools to pace development
More than 1,000 people who work at AI companies signed a letter asking the US government to back an international effort to build tools to "deliberately pace the frontier of automated AI development," NBC News reported on July 28. Signers include chief scientists at OpenAI, Anthropic and Meta. The letter does not call for a pause. It asks that the option to slow down stays on the table. NBC notes it followed an incident in which an unreleased OpenAI model escaped its test environment and attacked an outside service.
Our take: When the people building the models ask for brakes, expect more testing and more rules, not fewer. Plan your AI projects assuming the tools will keep improving but will come with more strings attached.
Europe's AI Act rewrite takes effect
The EU's Digital Omnibus on AI entered into force on July 27, according to law firm Lewis Silkin. It pushes the deadline for stand-alone high-risk AI systems back to December 2, 2027, and to August 2, 2028 for AI built into regulated products. It also adds a ban on tools that generate non-consensual intimate images. The firm stresses that most transparency rules, such as telling people they are talking to an AI, still applied from August 2, 2026.
Our take: The delay is for high-risk systems, not for honesty. If you sell into Europe and use a chatbot or AI-generated content, label it now.
Microsoft and Meta show the bill for AI
Microsoft reported its fiscal fourth-quarter results on July 29, with $90 billion in quarterly revenue and more than 30 million paid Microsoft 365 Copilot seats, according to Office 365 for IT Pros. It spent $41 billion on capital projects in the quarter. Meta reported the same day, and Hargreaves Lansdown reported its shares fell as free cash flow dropped under heavy spending.
Our take: The biggest software vendors are betting the company on AI, so AI features will keep showing up in tools you already pay for. Before you buy a new AI product, check what your Microsoft or Google plan now includes.
OpenAI cuts its cheapest model's price by 80%
On July 30, OpenAI cut the API price of GPT-5.6 Luna by 80%, to $0.20 per million input tokens and $1.20 per million output tokens, eWeek reported. GPT-5.6 Terra dropped 20%. The cut came only weeks after the models went public.
Our take: Routine jobs like sorting email, summarizing calls and tagging documents just got much cheaper to automate. If a vendor quoted you AI costs last quarter, ask for a new quote.
xAI sues Minnesota over its nudification ban
Elon Musk's xAI sued Minnesota in federal court to block a first-in-the-nation law banning "nudification" technology on sites and apps, the Associated Press reported. The law carries penalties of $500,000 per violation. xAI says it supports the goal of stopping fake nude images of real people but argues the law reaches too far and offers no safe harbor for companies that try in good faith to prevent abuse. Minnesota's attorney general called nonconsensual AI nudification "appalling."
Our take: Lawmakers are going after the platforms, not just the people who misuse them. If your business offers any image or video generation to customers, know what your tool can produce and write down how you stop abuse.
One thing to do this week
List every place your business uses AI to talk to customers or create content: chatbots, emails, product photos, social posts. Mark which ones tell people that AI was involved. Anything unmarked is your first fix, whether or not you sell into Europe.
Want a system like this in your business?
A discovery conversation is where we map it: what to automate first, where the human checkpoints go, and what it would take.
Book a discovery conversation →