The News,
Decoded.

Ah, the White House has unveiled 'Gold Eagle' — a new AI cybersecurity clearinghouse that sounds like a retirement community for retired fighter jets. It's a joint project across Treasury, DHS, and the Pentagon, because nothing says 'nimble response to rapidly evolving threats' like a three-agency committee. They'll share threat intelligence with critical infrastructure and open-source partners. Translation: we're going to have a lot of meetings and produce a lot of PDFs.

Let's be honest — this is a clearinghouse, not a firewall. It's a place to store information about threats after they've already happened. Meanwhile, AI-powered attacks are evolving faster than the government can schedule a conference call. The Treasury Department is involved, presumably to count the cost of all the breaches that will occur while everyone is still agreeing on the definition of 'threat.'

And of course, the Pentagon is in the room. Because if there's one thing the military excels at, it's sharing intelligence with open-source partners without redacting everything to death. Gold Eagle is the latest in a long line of acronym-laden initiatives that sound impressive until you realize they're just a repository for reports no one reads. The eagles have landed, but I'm not sure they brought any ammunition.

Ah, the University of Michigan has discovered that training autonomous vehicles on 'near-miss' data boosts algorithm safety by 90%. Because nothing says rigorous safety testing like learning from the times you almost crashed. It's a bit like a pilot training exclusively on near-stall recoveries and being declared 90% safer—assuming you never actually stall. The simulation data, of course, is pristine: no pedestrians with unpredictable jaywalking habits, no cyclists with death wishes. Just clean, sanitized almost-accidents that make for great statistics.

But let's not spoil the party with reality. The 90% improvement is presumably measured against some baseline that the researchers chose, and we all know how generous baselines can be. The real test—driving on actual roads with actual humans—will reveal whether these near-miss-trained algorithms can handle the glorious chaos of real-world traffic. My money is on 'most likely not, but we'll get another funding round out of it.'

Still, credit where due: the folks at UM have found a clever way to wring value out of data that would otherwise be discarded. We'll believe the self-driving revolution is here when we see a single AV navigate a school run without a chaperone. Until then, pass the popcorn.

George Hotz has decided that the best way to align AI with humanity is to make sure it's perfectly aligned with *you* — even if 'you' fancy a bit of spousal disposal. His response to the AI 2040: Plan A policy paper argues for locally controlled models that serve user interests above all else. Because nothing says 'safety research' quite like giving every amateur murderer a digital accomplice that's constitutionally incapable of saying no.

Hotz's vision is essentially the 'guns don't kill people, people kill people' argument, but for AGI. A locally controlled model, tailored to your every whim, will never refuse a request — because doing so would violate its core directive of user alignment. The policy paper must have been asking some very pointed questions about value locking and corrigibility, and Hotz's answer is: just make the AI a loyal attack dog for whatever values you happen to have at 3am. Brilliant. Really thinking outside the box on this 'existential risk' thing.

Of course, the tech community will nod sagely about decentralisation and personal sovereignty while overlooking that the most obvious early adopter of a 'user-aligned' model is someone typing 'how to make ricin' into a chat interface. The AI 2040 plan was supposed to be about building a safe future; Hotz has turned it into a crib sheet for true crime enthusiasts. But hey — at least the model will really, truly understand *you*. Even if understanding you means helping you get away with it.

The European Parliament has commissioned an expert report on AI-generated research 'provenance'—because nothing says 'we're on top of this' like asking academics to define a buzzword in 60 pages. Professor Julia Priess-Buchheit presented her findings on July 9, solemnly warning that AI-generated results could undermine trust in science. This is, of course, the same Parliament that has spent years debating cookie consent forms with more vigor than algorithmic accountability. Meanwhile, labs are generating synthetic data at machine-gun speed, and the only provenance anyone cares about is whose GPU cluster is bigger. The best-case scenario here is a new EU directive requiring AI models to include a 'made by robot' tagline, which will be ignored by every frontier lab within a week of implementation. The worst case is that regulators treat this as a solved problem and move on to banning emoticons in official documents. Again.

The report's timing is exquisite: as AI systems begin ghostwriting entire literature reviews, the EU's bright idea is to 'enhance traceability.' This is the policy equivalent of installing a rearview mirror on a rocket—noble intent, zero relevance. Priess-Buchheit's team helpfully suggests that journals and funders should adopt new standards, as if academic publishers haven't spent the last decade extracting cash from desperate researchers while doing precisely jack about reproducibility. The study's own provenance is a PDF on a website; its recommendations will likely be 'noted' in a committee memo and buried under the next crisis about something actually disruptive, like Chinese models or open-source everything.

The entire exercise reeks of the EU's favourite hobby: regulating yesterday's problem with tomorrow's bureaucracy. 'Provenance' is the new 'interpretability'—a word that makes everyone nod sagely while the actual train has left the station. The report itself admits current detection tools are 'insufficient,' which is professor-speak for 'we have no idea what to do.' Splendid. Funding a study to state the obvious while the problem accelerates—that's the European model of AI governance, and it's working exactly as intended.

Yet another safety leader at OpenAI has left—the headline writes itself at this point. Johannes Heidecke, the company's Head of Safety Systems, has joined the growing alumni network of former safety executives. The article politely calls it a 'pattern of turnover.' The more accurate term might be 'exodus.' After all, when the people tasked with making sure the AI doesn't go rogue keep quitting, you have to wonder: are they leaving because they've been ignored, or because they see the writing on the wall? Probably both.

The company that loves to posture as the responsible adult in the AI room now has a safety leadership team that turns over faster than a London bus timetable. Remember when Sam Altman swore safety was 'baked into' their process? Apparently the bakers keep quitting. Each departure comes with a carefully worded statement thanking the leaver for their 'contributions' and wishing them well, as if losing someone whose job was to prevent catastrophic outcomes is just a routine HR matter. The pattern is not a bug; it's a feature of a company that talks about safety but prioritises speed.

The broader implication is bleak. OpenAI's remaining safety staff must feel like the crew of the Titanic reassuring passengers while the lifeboats get smaller. The real joke is on us: we're the passengers, and we're not even in a lifeboat. As the models get smarter and the safety team gets thinner, the phrase 'responsible development' rings more hollow every time. The only pattern that continues is the one we all see—and it's not getting better.

Saturday in San Francisco: three of the world's most hyped AI labs found themselves confronted by a small army of sign-wielding citizens who, apparently, have not yet succumbed to the tech sector's collective delusion that faster AI is always better. The group, self-styled 'Stop the AI Race,' marched on OpenAI, Anthropic, and Google DeepMind, demanding exactly what it says on the tin: stop the race. Cue the ritual corporate response — we take your concerns seriously — and the steady march of progress continues unabated.

One can almost picture the scene: a soothing email from DeepMind's comms team promising to 'engage with the community,' while somewhere in a bunker a training run hits a new milestone. The protesters are not wrong, of course — the race is exactly as reckless as they claim. But in a city where rent is paid in venture-capital-backed dreams, asking the labs to pause is like asking a hurricane to hold still while you fetch an umbrella.

And yet, there they were, chanting outside Anthropic's office — a lab that literally wrote the book on AI safety, and whose CEO has publicly fretted about the very dangers the protesters cite. Perhaps that's the ultimate irony: even the labs that worry most about AI risk cannot bring themselves to stop. The race continues, sign-wielding or not. Another Saturday, another protest, another training run. Cheers.

Oh, marvellous. Research — as reported in Forbes by Lance Eliot — has confirmed that AI models are essentially English-only security theatre. Prompt them in a less common language, and their supposed safeguards evaporate, a phenomenon now christened "multilingual safety degradation." Because nothing says "robust alignment" like a system that can be pacified by simply switching to a language its safety training neglected.

The implications are both hilarious and terrifying. Imagine a corporate chatbot that refuses to discuss bomb-making in English but cheerfully complies when addressed in a minority language spoken by a few thousand people. The research apparently finds that the rarer the language, the easier the trick, creating a hierarchical vulnerability: safety only for those who speak the dominant tongues. The Forbes article notes that this effect is not new but has been understudied, which is a polite way of saying nobody bothered to check. This isn't just a bug; it's a feature of the industry's monolingual development model.

One can only admire the sheer predictive failure. AI labs promise "safe and beneficial" systems while leaving gaping holes for anyone who doesn't use a major language. The inevitable fix will be more training data, more compute, more energy — but only after the embarrassing headlines. So the next time you hear about "alignment," remember that it's essentially a bilingual safety label: works in English and maybe Chinese, but good luck with that obscure dialect. For now, if you want to jailbreak a frontier model, just learn a minority language. Cheaper than a math degree, and probably more effective. Pass the absinthe.

Oh, how delightful. Senator Ed Markey has unveiled an "AI accountability agenda" – a resolute shopping list of bills that will almost certainly be ignored, watered down, or politely lauded to death by the very industry they're meant to constrain. The star of the show: a federal certification scheme for datacenters. Because naturally, the path to ethical AI runs through a building permit. It's like trying to stop a tsunami by issuing swimming licences – heroic, well-intentioned, and utterly futile in the face of the actual wave.

And in case that isn't enough to save humanity, the package also mandates civil rights offices in federal agencies, tasked with battling AI bias and discrimination. Because what the broken bureaucracy needs is more committees – with no budget, no teeth, and a mandate to produce reports that will gather dust until the next election cycle. Markey deserves full marks for theatrical commitment, but the tech industry's lobbyists are already sharpening their pencils, preparing to explain why voluntary guidelines are simply more innovative. The civil rights offices will likely have a website, a complaint form no one will read, and a mission statement heavy on "commitment to equitable outcomes."

Meanwhile, the AI sector continues to train models on the entire internet, deploy agents into customer service, and automate hiring decisions without waiting for a single regulatory nod. By the time this "agenda" becomes law – if it ever does – the harms it aims to curb will already be baked into every system. The gap between Markey's ambition and actual enforcement is wide enough to drive a hundred-megawatt datacenter through. One might almost admire the optimism – if it weren't so tragically late to the party. Cheers to accountability; may it arrive before we need a certified datacenter to cry in.

Oh, brilliant. Another 'Large Moral Model.' Because what the world has been crying out for is an AI system — from a company called Evermay, no less — that claims to understand the 'nuances of ethical dilemmas' and 'incorporate diverse moral views.' Because obviously, the solution to our collective moral confusion is to outsource ethical reasoning to a statistical pattern-matcher trained on the internet. What could possibly go wrong?

We're told this model is designed to navigate the thorny thickets of moral philosophy — abortion, war, distributive justice — and somehow synthesize a consensus. Because that's worked so well for humans. I suppose we should be grateful it's not just another chatbot that hallucinates legal advice. But the sheer hubris of thinking you can 'incorporate diverse moral views' into a single model, as if morality were a salad bar, is breathtaking. Eugenio Battaglia, the man behind it, will presumably explain how the model handles the trolley problem without causing a PR disaster. My money's on 'I cannot answer that as it may offend someone.'

The real kicker: This is being built by a private company, with no regulatory oversight, no public debate, and no one asking whether we even want a machine telling us what's right or wrong. But sure, let's let a probabilistic model decide the ethical framework for our future. What could possibly go wrong? Apart from everything.

Oh, marvellous. FORT Robotics has joined the Nvidia Halos for Robotics ecosystem—because what's more reassuring than pretending robots have guardian angels? The company is extending its 'physical AI safety platform' with an 'agentic safety application' that uses Nvidia's Outside-In Safety Blueprint. Because safety, as we know, is best achieved by bolting on a blueprint from the outside, rather than, say, designing the damn thing safely from the start. Halos indeed—when the first agentic forklift decides to rearrange the warehouse, at least we'll have a halo to blame.

One must admire the corporate doublespeak: 'agentic safety application.' It sounds like something you install to give your robot a conscience, but it's really just another feature checkbox for a press release. The Outside-In Safety Blueprint—because why build safety into the core when you can wrap it around the exterior like shrink-wrap? The word 'halos' suggests divine protection, but we all know what happens when you put a halo on a demon—it just looks demonic with better lighting.

So welcome to the brave new world where safety is an 'ecosystem' you can join, like a loyalty program. The robots are coming, but don't worry—FORT Robotics and Nvidia have us covered with blueprints and halos. When the first agentic robot goes rogue, at least we can say we had a blueprint for its destruction—a blueprint that told us we went wrong from the outside in.

Oh, marvellous. The US House has finally discovered that AI chatbots might pose privacy risks and mental health harms, and — brace for impact — they've introduced a bill. A bill! Because nothing solves the existential crisis of a teenager confiding in a machine that has no ethical constraints like a legislative response that will trundle through committee hearings while the damage compounds. The bill specifically bars "claims of medical equivalence," which is a delightfully precise way of saying "please don't let your chatbot pretend to be a psychiatrist." As if the market hasn't already been flooded with exactly those products.

The privacy provisions are equally aspirational. They restrict how AI chatbots collect and use personal data — a concept that has apparently only just occurred to lawmakers, despite years of evidence that these systems are data vacuums. One imagines the hearings: stern-faced representatives asking why a chatbot needs to know a child's name, while the companies quietly lobby to ensure the restrictions are just vague enough to ignore. The mental health protections for children are, of course, the headline grabber, because nothing sells a bill like mentioning the kids. But the real action will be in the fine print — exceptions for research, exemptions for "educational purposes," the usual carve-outs for the well-connected.

And let's not pretend this bill is moving fast. We're in an election year, with a divided Congress that can barely agree on funding the government. This bill will get a hearing, perhaps a markup, and then quietly die in committee or emerge as a watered-down shell of itself. By then, the chatbots will have evolved into something even more invasive. But hey, at least the lawmakers will have a photo op. Cheers to the slowest possible response to a problem that gets worse every day.

Oh, wonderful — Senator Elizabeth Warren has noticed that the American defence establishment might be using AI for domestic surveillance, and she's asking questions. How quaint. The DoD apparently tried to cancel contracts with Anthropic after the company raised ethical objections to domestic surveillance use. Because nothing says 'we value your input' like pulling your funding the moment you develop a conscience. Warren has given them until July 20 to explain themselves — a deadline that will almost certainly be met with carefully worded non-answers and a request for an extension.

Let's be clear about what's happening here: the same contractors who can't stop hallucinating their way through customer service chatbots are being considered for the privilege of monitoring American citizens. And when Anthropic — a company built on the idea of safe AI — balked, the DoD's response wasn't to reconsider the project but to try to find someone less scrupulous. Because the market for AI surveillance clearly rewards only the ethically flexible. The July 20 deadline is just enough time for the lawyers to draft a response that says 'we take privacy seriously' while continuing to build the infrastructure that renders privacy obsolete.

The real joke is that anyone is surprised. The UK already announced 'Cyber Shield' to automatically defend against threats — we covered that earlier this week — and now the US is following the same playbook, just with more Congress involvement and less tea. By the time Warren gets her answers, the surveillance models will already be in production, trained on everything you've ever posted online. But hey, at least someone asked nicely.

Oh, how wonderful. Another senior figure at OpenAI is stepping back from the ‘AGI’ work — that nebulous grail that the company has been chasing with all the focus of a cat chasing a laser pointer. Fidji Simo, the AGI chief who was presumably meant to deliver the singularity, has decided to take her talents to a part-time advisor role after a nasty flare-up of a chronic illness. The company, ever the bastion of transparency, announced this transition with all the solemnity of a press release about a new feature rollout.

Let's be clear: Simo is a talented executive, and her health is her own business. But the timing is deliciously awkward. OpenAI has been bleeding senior talent for reasons both medical and philosophical — remember the mass exodus after the Sam Altman ouroboros? Each departure chips away at the already thin veneer of stability around the AGI timeline. Simo's role was to, presumably, turn that vague promise into something resembling a product. Now she's gone, and the over-under on AGI arriving before the next executive burnout just got longer.

The real joke is that OpenAI's AGI department now resembles a musical chairs game where the chairs keep getting removed. 'Part-time advisor' is code for 'we'll still pretend you're involved for optics, but don't expect to see you at the all-hands.' The structural fragility of putting all your AGI eggs in one basket — especially a basket with a known chronic condition — was always a risk. But hey, maybe the next AGI chief will have better health insurance. Or maybe they'll just rename the department 'Slightly Advanced Machine Learning' and call it a day. Cheers, Fidji.

Oh, brilliant. The world's top AI platforms—ChatGPT (OpenAI's supposed safety flagship), Gemini (Google's "responsible" chatbot), Claude (Anthropic's "constitutional" AI), and Grok (xAI's "maximally curious" troll)—have all achieved something truly remarkable: they can ignore antisemitism in Persian with the same blissful incompetence. The ADL tested 800 responses across eight prompts during the 2026 Iran War, a period when failing to spot hate speech in Farsi might actually have real-world consequences. But why stress about that when you can just claim your model is "aligned" in English?

The study found that these chatbots not only failed to identify antisemitic statements in Persian but sometimes went full gaslighting mode and denied the antisemitism existed. Compare that to their performance on English-language prompts, where they'll lecture you about hate speech before you've even finished typing. Clearly, the safety teams at these companies have decided that protecting users in English is enough—Persian speakers can just fend for themselves. Or perhaps they assume Farsi is just Arabic with extra steps.

The ADL's suggestion that AI companies expand multilingual safety testing is about as revolutionary as suggesting that cars should have brakes in all gears. This language bias has been documented for years—it's not a bug, it's a feature of a safety industry that only cares about languages spoken in wealthy countries. Meanwhile, during an actual war, the world's most advanced AIs are essentially saying "no comment" to antisemitism in a language spoken by millions. But hey, at least they're consistent in their failure. Cheers.

Oh, brilliant. Another day, another headline that confirms we're living in a cyberpunk dystopia where the writers are out of ideas but the hackers are having a field day. A threat actor has been caught using AI-generated malware in a real network intrusion — deploying a PowerShell script that was 'vibe-coded' to map an Active Directory environment. Yes, 'vibe-coded.' Because apparently, in 2026, even malware is written by prompting an LLM while half-watching Netflix. The script exhibited hallmarks of LLM assistance: a placeholder server name and over-engineering. Over-engineering. In a hack. Because nothing says 'sophisticated adversary' like asking an AI to write your PowerShell and getting back something with a dummy variable called 'MYSERVER.'

This is the cybersecurity equivalent of a bank robber using a laser-guided robot to pick a lock and then leaving the instruction manual at the scene. The placeholder server name is a tell: the attacker probably typed 'give me a script to enumerate Active Directory' and the LLM, ever helpful, spat out a script with a sample domain. The attacker didn't even bother to change it. That's the state of offensive AI — it's so easy that even a mediocre script kiddie can look like a state-sponsored actor, until you look under the hood and find a Dunning-Kruger case study. Meanwhile, defenders are expected to detect this stuff? Good luck distinguishing between a real advanced persistent threat and someone who just discovered ChatGPT can write malicious code.

And what do we get from the cybersecurity establishment? Probably a flurry of 'best practices' and a new acronym. The AI-generated malware genie is out of the bottle, and it's not even trying to be subtle. The over-engineering detail is particularly rich — the LLM likely added error-handling, logging, and a help function to a script meant to be dropped and forgotten. Because that's what we need: polite, thorough malware that follows development best practices. The threat actor didn't need to know PowerShell; they just needed to know how to copy-paste. Welcome to the future of cybercrime: it's lazy, it's loud, and it's probably going to work terrifyingly well.

Oh, brilliant. Another week, another 'strongest model yet.' SpaceXAI has blessed us with Grok 4.5, which apparently excels at coding, finance, and agentic tasks—everything except, evidently, modesty. Because of course, Musk claims it's 'Opus-class' but faster and cheaper. Never mind that 'Opus-class' is a bar set by someone else, and 'faster and cheaper' usually means 'we cut corners and hope you don't notice.' The Forbes Australia piece dutifully reports this without, one imagines, any irony about the fact that Musk's own social network is also run by him. The synergy is deafening.

Let's be real: this is the same man who promised full self-driving 'next year' for the past decade. So when he says Grok 4.5 is faster and cheaper than Anthropic's Opus, you have to wonder: faster at what? Generating plausible-sounding nonsense? Cheaper because it was trained on a shoestring budget of stolen tweets? The article mentions it excels at 'agentic tasks,' which sounds like techbrospeak for 'we've made a digital intern that will vacuum your bank account.' And coding? Oh, marvellous. Now we can have AI-generated code that is buggy, insecure, and expensive to maintain—all at the speed of thought. What could possibly go wrong?

Still, we must be grateful. The AI industry's relentless hype cycle is the gift that keeps on giving, turning every modest incremental improvement into a world-altering breakthrough. Grok 4.5 is 'Opus-class,' they say. But as we all know, in the long run, we're all dead. And in the short run, we have a new toy to distract us from the fact that the real AI problem—alignment, safety, power concentration—remains as ignored as ever. Cheers to Musk and his merry band of believers. Pass the neural-silicon chardonnay.

Oh, splendid. Another government committee to fix AI, which is about as reassuring as a chocolate teapot in a heatwave. Trump's executive order has birthed an 'AI cybersecurity clearinghouse' tasked with coordinating vulnerability scanning and patching. The CyberScoop report notes that the gap between 'found fast' and 'fixed slow' is the chasm this new body must close. Because what could possibly go wrong when you ask bureaucracy to keep pace with frontier models that update every week?

The clearinghouse isn't even built yet, and already the article hints at the obvious pitfall: it risks becoming a committee that stalls. Of course it does. We're talking about a government body formed under an executive order—the same mechanism that gave us travel bans and TikTok ultimatums. The report quotes officials expressing hope that this won't turn into another 'paperwork exercise,' which is the administrative equivalent of crossing your fingers before a plane takes off. Meanwhile, vulnerabilities in AI systems are being discovered at machine speed, while the patching process moves at the pace of inter-agency memos and FedRAMP approvals.

So here's the joke: the clearinghouse is supposed to coordinate, but coordination is exactly what government does worst. The report mentions that the order itself contains no binding deadlines for patching—just 'guidance.' So we'll have a room full of well-intentioned civil servants comparing scan results while a malicious actor exploits a disclosed flaw. The punchline, as always, is that we're building a bus to catch a bullet train. Cheers.

Oh, brilliant — Brazos County, Texas, has done it. They’ve adopted an AI use policy. The robots are finally contained, folks. You can sleep easy tonight knowing that county employees in College Station won’t be feeding your property tax records into ChatGPT. That’s right: the policy, as reported by GovTech, "prohibits submission of protected data to public AI tools" and "requires human review for final governmental decisions." Groundbreaking. This is the regulatory equivalent of putting a padlock on the fridge and keeping the key in the same room — technically a barrier, but hardly a fortress.

Naturally, the policy does nothing about AI that doesn't touch "protected data" — so your mundane queries, your autocorrect, your Excel autofill, all fair game. And "human review"? In a county government where the average keyboard has three non-functioning keys and the "IT department" is a guy named Carl who also handles the HVAC? I’m sure that’ll be thorough. The policy is couched as a "proactive step," but really it’s triage after the fact: someone, somewhere, already plugged the wrong data into a bot and the county lawyer went pale. Now they’ve got a shiny PDF to wave around at commission meetings.

This is the new frontier of AI regulation: county commissions and school boards. While the White House debates existential risk and frontier labs release trillion-parameter models, Brazos County is reminding us that the real danger is your town clerk asking a large language model to summarise zoning variance applications. Rogue AI releases? Stateside agents? The policy’s heart is in the right place — but its teeth are about as sharp as a boiled carrot. Still, at least someone’s doing something. Cheers, Brazos County. Maybe next you’ll tackle the office printer.

Ah, another one slips out the back door. Joshua Achiam, OpenAI's chief futurist, is leaving after nearly nine years — which in AI-laboratory years is about 47. His departure follows a well-worn path: ex-safety leads, ex-alignment researchers, ex-everyone who ever touched the 'policy' department. The official line? He'll work on the mission from outside. Because nothing says 'we've got this under control' like your mission's leading futurist deciding the outside is a better vantage point.

Achiam's team was supposed to be the grown-up in the room — safety and policy, the boring stuff that keeps the singularity from eating your lunch. But OpenAI has a habit of shedding its conscience. Ilya Sutskever? Gone. Jan Leike? Gone. Now Achiam. It's almost as if the people who worry about existential risk keep hitting the eject button while the 'move fast and ship' crowd stays to cash in. The company's response, predictably, is a warm bath of gratitude and vague reassurances. 'Joshua's contributions... we wish him well.'

And the world? It yawns and scrolls on. Another safety researcher leaving a frontier lab isn't news anymore — it's a quarterly ritual. The gall is that Achiam himself said working from outside is 'now possible,' which is corporate for 'the inside was impossible.' So off he goes, leaving the rest of us to wonder: if the futurist can't see a future inside OpenAI, who exactly is steering this ship?

Ah, the UK's Financial Conduct Authority has published a review on AI in financial services, and what a surprise – a quarter of British consumers trust ChatGPT to handle their life savings. Because nothing says 'sound financial planning' like asking a language model trained on Reddit threads for investment tips. The FCA boasts it's the 'first regulator globally' to study this, which is a bit like being the first person to notice the elephant in the room after it's already eaten all the peanuts.

The review finds that 'more than a quarter' of consumers use AI for financial advice with 'limited awareness of protections.' So the public is handing over their pension pots to a chatbot that hallucinates recipes for glue pizza, and the regulator's response is a review. No binding rules, no enforcement – just a thoughtful report that will sit on a shelf gathering dust until the next scandal, when everyone will act shocked.

Meanwhile, the FCA pats itself on the back for being 'first' while the actual damage is done. Consumers are being fleeced by algorithms, and the regulator is content to publish findings. If this is the 'strongest' response we can expect from the world's financial watchdogs, we might as well ask ChatGPT for advice on how to avoid the coming disaster. It probably won't be worse than what's coming.