UK Regulators Just Gave Publishers a Way to Say No to AI Search
Today’s AI news sits at an interesting crossroads: the people and companies who create content are starting to push back against the systems that consume it. From a landmark UK regulatory ruling to questions about what’s real on Amazon, and a new White House order on AI oversight, June 4th is a day about control — who has it, and who’s asking for more of it.
UK Publishers Can Now Opt Out of Google’s AI Search Results
First, a quick definition: AI-generated search results (sometimes called “AI Overviews”) are the summaries that Google now shows at the top of many search pages — written by an AI instead of linking directly to websites. Publishers have long complained that these summaries pull from their work without sending readers to their actual pages.
That’s now changing. UK regulators at the Competition and Markets Authority have ordered Google to build a tool giving website publishers the ability to opt out of having their content used in these AI-generated summaries. The ruling also requires Google to provide clearer attribution — meaning visible links back to the original sources when AI search features do use publisher content. This affects any publisher whose content appears in Google search in the UK, from major newspapers to small blogs.
In plain English: imagine you run a recipe website. Right now, Google might summarize your chocolate cake recipe at the top of search results, and many users never click through to your site. Under this new requirement, you’d have the power to say “don’t use my content that way” — and if Google does use it, your site must be clearly credited with a working link. It’s a meaningful shift in how the relationship between search engines and content creators is structured.
Why this matters: This is one of the first times a major regulator has forced a concrete, usable tool into the hands of publishers — not just a vague promise of fairness. It could set a precedent that other countries follow.
“UK regulators are requiring Google offer a tool allowing website publishers to opt-out of generative AI search features.”
You can read the full story at TechCrunch, with additional coverage at Ars Technica and Courthouse News.
Amazon Is Showing AI-Generated Product Images in Search — and They’re Not Real Products
AI-generated images are pictures created entirely by an artificial intelligence, with no photographer or physical object involved. Amazon has started displaying these in its shopping search results to represent broad product categories when actual seller photos aren’t available.
So if you search for “gray linen throw pillow,” you might see a polished, realistic-looking image of one — but that image doesn’t correspond to any product you can actually buy. It’s meant to illustrate the idea of what you’re searching for, helping shoppers visualize a category before real listings appear. Amazon says these images are labeled and are intended as a browsing guide, not a product listing.
Here’s where it gets a little complicated for everyday shoppers: realistic AI images in a shopping context could easily be mistaken for real product photos. The line between “this is a helpful illustration” and “this is what you’re buying” needs to be very clearly drawn — and right now, how well Amazon draws that line is something worth watching. For sellers, it also raises questions about whether AI placeholder images might affect how their actual products get noticed.
Why this matters: This is an early example of AI-generated visuals entering spaces where people are making real purchasing decisions, and it sets a precedent for how much visual “assistance” platforms can offer before it starts to mislead.
“AI-generated images that match your search queries to help guide users.”
Full details at TechCrunch and CNET.
Trump Signs Executive Order Asking AI Companies to Voluntarily Submit Models for Review
An executive order is a directive from the President that instructs government agencies on how to act — it doesn’t require Congressional approval. President Trump signed one this week directing AI companies to voluntarily submit their most advanced AI models (the core software systems behind tools like chatbots and image generators) to the federal government for review before releasing them to the public.
The administration is framing this as a national security measure. The idea is that government reviewers would assess whether a powerful new AI poses risks to critical infrastructure (think: power grids, financial systems, or military applications) before it reaches the public. Crucially, the word “voluntarily” is doing a lot of work here — companies aren’t legally required to comply.
For everyday people, this order doesn’t immediately change what AI tools are available. But it signals that Washington is thinking more seriously about what happens when extremely capable AI systems are released without any outside check. Whether voluntary review is meaningful oversight, or just a gesture, will depend entirely on whether companies actually participate.
Why this matters: This is the first significant federal action asking for a look inside powerful AI systems before release — even if it lacks enforcement teeth for now.
Coverage from The New York Times, The Guardian, and CNN.
Also Happening in AI
It was a busy day beyond the headlines. Google quietly released Gemma 4 12B, a new open multimodal model (an AI that can understand both text and images) with 12 billion parameters designed to run on consumer hardware — good news for developers who want powerful AI without cloud costs. Microsoft launched Scout, a new AI personal assistant built into Microsoft 365 apps like Outlook and Teams. Observability startup Coralogix raised $200 million at a $1.6 billion valuation, per TechCrunch, betting that businesses will need dedicated tools to monitor and manage the AI agents they’re increasingly deploying. And Wired has a detailed look at how Trump’s AI executive order came together after an earlier version was shelved.
What to Watch
The common thread today is accountability — regulators, governments, and consumers are all starting to ask who’s responsible when AI gets something wrong, takes something without asking, or releases something risky. The UK’s publisher ruling and Trump’s executive order, despite coming from very different political contexts, both reflect the same underlying pressure: AI is powerful enough now that the “move fast” era is giving way to something more deliberate. Watch for other countries to respond to the UK ruling in particular — it may be the template that shapes how AI search works globally.