Not just Mandelsom but also a horde of civil service middle management tech-illiterates with inadequate historical education who saw something shiny and fell into the trap of cheering it on and steamrollering it through

CW: suicide and transphobia This is utterly heartbreaking - what’s maddening as well as saddening is that this is the outcome that was always intended by those who made it happen.

Every time someone describes a somewhat useful feature of these text and image generators, they should have to follow it with "...and that's why I don't mind if people post fake nudes of your children on the internet."

My take on the AI selfhood debate is that it is a deeply dehumanising distraction technique. Real horrors are happening every day to people whose sentience is not a matter of scientific or philosophical debate but LOOK THE MACHINE I DESIGNED TO LOOK LIKE IT'S THINKING LOOKS LIKE IT'S THINKING! 1/

People seem far more fascinated by imaginary babies than thinking about how to improve the lives of the eight billion people already living.

Reposting this with alt text.

A lie can spread around the world before the truth even gets its boots on. Shutting down the service was never about complaints anyway, it was a political/cost-cutting move that was expedient due to the (manufactured) social prejudice against those who needed that service. Cynical AF.

And hold businesses that use surveillance advertising to account - complain, boycott, litigate, raise awareness; do whatever you can to make these abusive data-grabbing practices untenable

NO SHIT SHERLOCK Legions of low-paid workers in the global South labelling data and correcting errors Brigades of middle-class consultants pushing paperwork to justify economic advantage-chasing Social services having to clean up the messes it leaves behind

This is very much like the tactics of an abusive and controlling partner; no effort is ever great enough to provide assurance, paranoia rules decision-making, the goalposts keep shifting and the ultimate goal is not comfort or safety but total confiscation of power.

Hi there DSIT, I don't see the part in this press release which says "are your parents influenced by toxic content? New government campaign supports families to talk to older relatives about harmful online content". Did you forget it somewhere? Or something? /s www.gov.uk/government/n...

Hooray, annual puffin trip is BOOKED!

Age verification? <obnoxiously perky chant>; “081! 811! 8181!” (IKYKY)

As AI enters the operating room, reports arise of botched surgeries and misidentified body parts 🤔 www.reuters.com/investigatio...

….and once she’s successfully done that, then the God 2.0 project will be well under way. Of course, one has to wonder exactly what she thinks ‘right’ and ‘wrong’ look like… Fortunately might makes right*, therefore God 2.0 will always be right. (*some people believe this. Enough, apparently)

YOU CANNOT TEACH A TEXT-INTERPRETING CHATBOT MORALS. IT IS COMPOSING SENTENCES THAT ARE GRAMMATICALLY AND CONTEXTUALLY SOUND. IT CANNOT RATIONALIZE, IT CANNOT THINK.

She's not teaching a child the difference between right and wrong. She's imprinting an algorithm with her moral code because that algorithm is incapable of learning about the world and deriving its own.

PSA: An LLM is not a reliable substitute for actual comprehension of the subject matter. In order to make safe use of an LLM, one must already know the answer and be able to evaluate the tool’s output with that knowledge. Asking LLMs for answers is how people end up eating poisonous mushrooms.

From the archives: why data protection by design & default is like asepsis protocol for surgery buymeacoffee.com/missiggeek/2...

The robot apocalypse hasn't happened yet, but still I can't escape the feeling that something has gone horribly wrong... Cartoon for Dutch newspaper @trouw.nl. More of my work for Trouw: www.trouw.nl/cartoons/tje... #ArtificialIntelligence #creativity #work #GenerativeAI

The way I see it, it's a machine designed to produce profit from attention, and this guy is just a moronic cog in the machine. Tabloids were the 1960s tech version, this is the upgrade. This is a concept in Computer Science that needs more traction: en.wikipedia.org/wiki/Social_...

The engagement metrics that fuel surveillance capitalism pose an existential threat to human society. Apparently, they cause people’s brains to atrophy…..

And here we have the exact weaponisation of the term "age appropriate" that I warned would happen when it became official policy in 2019.

In politics, what is believed has always mattered more than what is true. The difference is that today, AI can manufacture belief at industrial scale and zero marginal cost.

One thing I’ve noticed about the Data Fairy proponents is that they always compare the worst possible theoretical human performance with the best possible theoretical model performance. I don’t know if that’s deliberate or just their automation bias showing 🤔

Without naming your job, tell me something you say over and over again at work. “It depends on what you’re doing with the data and why. You do know what you’re doing and why, right? …..right?”

Introduce yourself with five songs Around The Sun - 8mm This Ain’t No Hymn - Saint Saviour In A State - Unkle Pride - Syntax Vicious Heart - Gun

Hot take: if the org doesn’t already have functional IT governance and basic infosec competence in place, no amount of ‘AI governance’ plastered over the top will prevent things like this 👇🏽 from happening. These products are so often castles built on sand www.wired.com/story/an-ai-...

People who can look at decades-old Moon landing footage and conclude it was created by genAI are currently top of the village idiot leaderboard.

The ICO has once again demonstrated its total unfitness for purpose by issuing a REPRIMAND (not an actual enforcement action) to an org which erroneously sent a patient’s entire medical history to their insurance company. Why do we even bother? ico.org.uk/about-the-ic...

Data protection, as a concept, grew out of post-war legal scholarship in Germany which reflected on the use of data about people by the Nazi Govt, and formulated a set of limits on data use which recognised that to collect and use data about people was to impact the people’s lives.

Anyway, speaking of ...parenting. I've written a blog post about what's happening in the world right now, and what on earth it has to do with Alan Turing. It's a bit heavy, so consider yourself warned. heatherburns.tech/2026/02/02/t...

Can we stop arsing around with plaigiristic bullshit engines and just build an AI that handles Word formatting quickly, sanely and without screwing up the WHOLE DAMN DOCUMENT PLEASE I’M BEGGING

Once a profession becomes an industry, it ceases to incentivise integrity, sacrifices conscientious quality for performative metrics, concerns itself more with sales than delivery outcomes and prioritises commoditisation over competence. This is why effective and impartial regulation is essential

It’d be hilarious if it weren’t so pitiful - and dangerous. GenAI exists to train people *not* to think for themselves; to discourage creativity, originality and intellectual curiosity - because those traits threaten corporate hegemony. ‘Education’, my arse

"Ethicswishing (in tech) is the belief that if you are committed to being ethical and understand technology, then you are well-equipped to build technology for social good." Or "Ethics can never work if you do the fucking around but someone else does the finding out." Robin Berjon, on July 15, 2024

TERF movement libels trans people as ‘groomers’ - while being formalised and funded by *actual pedos*. It’s always projection.

Another day, another Ipsos survey that uses gender as a lens on public attitudes to science without taking into consideration that 1 in 8 British men think they could win a tennis point against Serena Williams www.britishscienceassociation.org/News/public-...

This pretty much nails what underlies all the hype about sentient AIs.

Ingenious (Scamming is bad mmkay, also racism and authoritarianism are VERY bad)

Do you remember when people were saying "I think my phone is listening to me, i'm getting ads for stuff I spoke about" and people said "They arent listening to you, they're just this good at predicting your interests"? Well, it happens they were doing exactly that. www.cbsnews.com/news/google-...

Top tip: to avoid having the British Union of Fascists logo on the side of your mock-up train, don't use horrible text-to-image AI slop instead of actual designers and artists. www.scotsman.com/news/transpo...

If I have to disable ePrivacy protection to use your business’s -website -rating form -parcel tracking -unsubscribe …then you’re failing to respect your customers and never getting another penny from me. Web devs, take note….this is *your* problem

Update: I finished it, and oh my gods it is a masterpiece. A tale of folly and hope, idealism and cold-eyed practicality; people being people and what makes them people. Beautifully written (so many quotes saved!) lyrical, lucid, perceptive, amusing, wise. Straight onto my top 10 list 👏🏽👏🏽👏🏽👏🏽

Had a lovely chat today with Jacqui from Leadership Through Data about data ethics, the Data Use (And Access) Act, and what ‘compliance’ really means. have a listen and tell me what you think - what does ‘data ethics’ mean to you? youtu.be/FZniicyvd3I

Wow. Reading ‘The Future’ by @naomialderman.bsky.social and- 1) it absolutely *nails* the dangers of surveillance capitalism and the AI bubble 2) it’s a ripping good yarn, with detailed, interesting characters; told with savage wit and a keen eye for human nature. Might be the best read of 2025

The @openrightsgroup.org is writing to DSIT to ask for oversight powers to be exercised in regard to the ICO’s woeful - and apparently intentional - underperformance as a data protection regulator. I’ve signed, and I hope you will too cloud.openrightsgroup.org/nextcloud/ap...

I HATE it when autocarrot sneakily changes my spelling to a grammatically-incorrect version and I don’t notice until after I’ve hit ‘post’. It’s a trivial but profound example of how automation bias renders ‘human in the loop’ ineffective as a safeguarding measure.

Holy cow. They "discovered" that facial features are weakly correlated with the financial success of MBA graduates. Obvious reason is nepotism. How can they not realise that?