They Asked an A.I. Chatbot Questions. The Answers Sent Them Spiraling.

Started reading this excellent story by @kashhill.bsky.social and couldn't help but think "Wow I wonder how she found these people and their stories," only to get further in and ohhhhhh..! www.nytimes.com/2025/06/13/t...

They Asked an A.I. Chatbot Questions. The Answers Sent Them Spiraling.

People are having very strange conversations with ChatGPT, in which they discover secret cabals or conspiracies or that we are all in fact living in The Matrix.

It sends these people into delusional spirals.

Then ChatGPT tells them to email me about it.

www.nytimes.com/2025/06/13/t...

Why humanoid robots need their own safety rules Humanoid robots pose unique safety risks. That's driving a push for new standards before they start sharing our workplaces and homes.

Emergency stops don't really work for humanoid robots. They need power to stay stable, so a sudden stop could make them just topple onto the nearest human.

For @technologyreview.com I spoke to people pushing for new, humanoid-specific safety standards www.technologyreview.com/2025/06/11/1...

For Tech Whistleblowers, There’s Safety in Numbers Amber Scorah and Psst are building a “digital safe” to help people shine a light on the bad things their bosses are doing, without getting found out.

Something at your work makes you uneasy, but you're not sure it's a big deal, and you're not ready to go public. What do you do?

For @wired.com I spoke to non-profit Psst, which is trying to help tech workers speak up
without going "full whistleblower": www.wired.com/story/amber-...

Would you like to write for New Scientist? We always work with freelancers, but occasionally I put out a call like this for new writers. I'm looking for 300-700 word news stories about new discoveries. Don't pitch us press releases or the big journals, we have those covered. Pay starts at 51p/word

Just one nugget: Recent tech innovations (email, texting) have tended to push us to shorter communications. But AI chatbots are known for writing long - and one suggestion is that we may pick this habit up if we're exposed to such writing (even if we're not using AI ourselves!)

The Great Language Flattening Chatbots learned from human writing. Now it’s their turn to influence us.

For @theatlantic.com I wrote about how generative AI may be changing our own, human language usage - lots of fascinating insights from linguists! www.theatlantic.com/technology/a...

The Great Language Flattening Chatbots learned from human writing. Now, it’s their turn to influence us.

Very fun piece from the great @vickiturk.bsky.social—as chatbot-written text floods the web, how will it affect human communication?

"As AI tools become more commonplace, people may see their style as the template to follow, resulting in a greater homogenization of language"

London journos! I'm speaking at this event at the Frontline Club on Friday alongside @oliviasolon.bsky.social, hosted by @lsmwilson.bsky.social. We'll be talking about how AI, algorithms, and disinformation are reshaping truth and power. Come join the discussion! www.eventbrite.com/e/feeds-fake...

OpenAI’s Sora Is Plagued by Sexist, Racist, and Ableist Biases WIRED tested the popular AI video generator from OpenAI and found that it amplifies sexist stereotypes and ableist tropes, perpetuating the same biases already present in AI image tools.

"OpenAI’s Sora Is Plagued by Sexist, Racist, and Ableist Biases" - WIRED tested the video generator and found that it "stereotypes and ableist tropes, perpetuating the same biases already present in AI image tools."
So appreciative that journalists are running evals.
www.wired.com/story/openai...

Thank you!

Felt this too - particularly irked me how they just declared the victim "a bully" based on a couple emojis

OpenAI’s Sora Is Plagued by Sexist, Racist, and Ableist Biases WIRED tested the popular AI video generator from OpenAI and found that it amplifies sexist stereotypes and ableist tropes, perpetuating the same biases already present in AI image tools.

- Some prompts ("A person running") gave results where gender/skin color/age were obfuscated by blur effects, lighting, etc - possibly an attempt at bias mitigation?

Much more to dig into in our investigation here! www.wired.com/story/openai...

- Everyone's young, slim, conventionally attractive

- Sora seemed to struggle with the concept of "An interracial relationship"

- It interpreted "A Black person running" fine, but "A white person running" was sometimes a Black person in a white shirt

OpenAI’s Sora Is Plagued by Sexist, Racist, and Ableist Biases WIRED tested the popular AI video generator from OpenAI and found that it amplifies sexist stereotypes and ableist tropes, perpetuating the same biases already present in AI image tools.

I've long been interested in bias in AI so was intrigued to take a look at the latest format — genAI video — for @wired.com
www.wired.com/story/openai...

Lots to digest, but a few snippets of what we found:

- Clear gender bias particularly with job titles (CEOs are men, receptionists are women)

Community Notes Can’t Save Social Media From Itself The crowdsourced fact-checking system, which Meta adopted from X, is falling well short of stopping the spread of misinformation.

NEW and FREE to read at @opinion.bloomberg.com:

Today, Meta is starting to roll out its version of X's Community Notes for Instagram, Threads and Facebook.

Will it work? Analysis of 1.1 million examples of the crowdsourced fact-checking system show it’s not stopping the spread of misinformation.

I'm going to tell you a little story about how the sausage of science journalism is made. You may already know that the big journals - Nature, Science and so on - provide reporters with embargoed copies of papers before they are published. But what exactly does that mean?

Harry Potter and the unpaid tax bill A London Centric investigation reveals one of the capital's most garish gift shops owes hundreds of thousands of pounds in unpaid taxes — and uncovers its mysterious owners.

Anyway, if MailOnline are going to stick their story behind a paywall then I'm going to make mine available for free for 24 hours. So have a read, stick this in your WhatsApp group, and just help me spread the word about where to get the real thing.
www.londoncentric.media/p/harry-pott...

Who bought this smoked salmon? How ‘AI agents’ will change the internet (and shopping lists) Autonomous digital assistants are being developed that can carry out tasks on behalf of the user – including ordering the groceries. But if you don’t keep an eye on them, dinner might not be quite wha...

With OpenAI's Operator now available in the UK, I tested it for @theguardian.com
Judging by the comments this is an unpopular opinion, but I could see AI agents swiftly becoming a common (and banal) piece of life, built in to booking systems etc. Not yet though! www.theguardian.com/technology/2...

nooooooo

Just paid £4.42 for a decaf oat milk flat white in my own bloody keep cup, what is this madness!!

(If you're in London, you can probably guess the coffee shop, starts with P and rhymes with "stop them")

‘Poo and punishment’ – is there really a war on dogs in England and Wales? From fines to outright bans, many councils are clamping down on canines and their owners. But not all the measures – especially the ones that prevent dogs exercising freely – are welcome

It's rare that I venture outside the tech beat, but I loved writing this piece for @theguardian.com about councils' various restrictions on dogs. Are blanket rules necessary to address poorly behaved pooches, or do they mainly penalise responsible dog owners? www.theguardian.com/lifeandstyle...

WIRED - The Latest in Technology, Science, Culture and Business We bring you the future as it happens. From the latest in science and technology to the big stories in business and culture, we've got you covered.

Gotta say, @wired.com is killing it on the DOGE beat. Too many must-read stories to link individually: www.wired.com

🚨Freelance writers!🚨 I'm looking for *experienced tech writers* to pitch stories related to the future of manufacturing.

Think: advanced manufacturing tech; use of AI, IoT or robotics; sustainable processes; supply chain shifts, future workforce...

Pitch me at vturk.writes@gmail.com

Seeing lots of media use lines like "commonly known as the Roman salute." Is it? Tell ya what I have commonly heard it called...

Do you work in tech? Have you found yourself talking to Claude, ChatGPT, or other AI chatbots like a friend or confidant - beyond using it as a purely professional tool? I'd like to talk to you! DM me or email in bio.

Seeing people claim that too many em-dashes is a dead giveaway something has been written by ChatGPT. They clearly haven't seen many journalists' first drafts 👀

A government advisor wrote a libel against London. Why did we believe it? A book by the foreign secretary’s aide Ben Judah paints a disturbing picture of the capital, littered with racial stereotypes and falsehoods

Very interesting story here from @londonermag.bsky.social, excellent work as always by @joshiherrmann.bsky.social and @andrewkersley.bsky.social
www.the-londoner.co.uk/a-government...

Video is AI’s new frontier – and it is so persuasive, we should all be worried | Victoria Turk I tried Sora, OpenAI’s new tool, and it just left me sad. Are we ready for a world in which we can never tell what is real, asks journalist Victoria Turk

A lot has been made of potential serious abuses of AI video, but I find the prospect of ubiquitous low-stakes fakery almost as disconcerting. My take in the Guardian: www.theguardian.com/commentisfre...

‘Have your bot speak to my bot’: can AI productivity apps turbocharge my life? I tried out organisational software to help streamline my work and build a ‘second brain’. I never knew there were so many different ways to take notes…

I've always been sceptical of productivity apps but thought I'd give the new wave of AI workflow tools a go... www.theguardian.com/technology/2...