In this month’s Roundup: A new report by the European Commission finds that X is the worst platform for disinformation, is this the creepiest AI to date? And Jony Ive’s back.


According to a new report, X is the largest source of mis- and disinformation in the EU, writes Lisa O’Carroll for The Guardian.

Conducted by the European Commission to assess compliance with the newly implemented Digital Services Act (DSA), the report concluded that X scored worse in Disinformation Discoverability, Ratio of Disinformation Actors and Relative Post Engagement where a post was found to contain disinformation. When looking at Absolute Post Engagement for disinformation, TikTok was by far the worst.

A code of conduct is in place to help companies prepare for the changes enforced by the DSA. Most of the large tech firms are signatories, however X withdrew from the code earlier this year.

Meta is the latest to enter the AI chatbot race, writes James Clayton for the BBC. The chatbots will be deployed on Messenger and are designed to have “personality”.

That “personality” will be supplied by a range of celebs that have signed on to lend their voices to the chatbots. As per the article, Snoop Dogg, Tom Brady and Kendall Jenner will all play “characters,” in Brady’s case “a wisecracking sports debater” named “Bru”.


Creepy uses for AI are bountiful, but Clearview AI’s facial recognition tech takes some beating.

For those not in the know, Clearview AI scraped a host of platforms and collected people’s faces, all without consent, natch. These faces were then used to create a facial recognition app that can scan the internet for a given person. In total, the company has now harvested 30 billion images.

In this extract from her new book, excerpted in Rolling Stone, Kashmir Hill outlines the beginnings of Clearview AI and the group of very wealthy people who were early fans.

One of the attendant hazards of a life lived online is hacking. But besides hacking into emails, bank accounts and the like, there is another, slightly stranger hacking target: Facebook accounts.

A group of hackers in Vietnam have been targeting Facebook accounts, taking control of them and then selling them. As one ex-hacker put it: “A new Facebook account has no value at all, but an old Facebook account is so valuable on the market.”

Amanda Florian’s article at Vox reports on who these hackers are, how these stolen accounts can be used and how you can protect yourself.


If you watched that WeWork documentary, you may recall SoftBank money-man Masayoshi Son, who poured billions of dollars into the company before realizing his error and offering CEO Adam Neumann over a billion dollars to share any space other than the boardroom.

Well, he’s back. Fresh off announcing an annual loss of $32 billion from his Vision Fund, Masayoshi Son looks set to fund a new OpenAI x Jony Ive collab to produce a “more natural and intuitive user experience” with AI.

According to Jess Weatherbed at the Verge, discussions are at an early stage, although there’s apparently been a lot of “brainstorming”.

Google will now let you refuse permission for your website to be used to train its AIs, writes TechCrunch’s Devin Coldewey. The cynics amongst you may note that this wasn’t an option before it developed its range of AI tools.

At the same time, a host of authors have begun legal proceedings against OpenAI for copyright infringement, claiming that their works had been ingested to train data without their consent.