Artificial Intelligence - General News

[Graham1973]

Commentary video on the recent spate of 'autocompletes for life' going rogue

View: https://www.youtube.com/watch?v=f-Fx8dJPY5M

 

[Mr.T]

The technical term for what's being described is 'Ponzi Scheme'....



Top economist warns that the AI math doesn’t make sense: 'Profits are currently being funded by investors rather than earned from customers'

At least now they can pretend the whole circle jerk was fu*ked up by Chinese ,

 

[shin_getter]

Now, for advanced Chinese harness tech:
https://github.com/tanweai/pua

An AI Coding Agent skill plugin that uses corporate PUA rhetoric (Chinese version) / PIP — Performance Improvement Plan (English version) from Chinese & Western tech giants to force AI to exhaust every possible solution before giving up. Supports Claude Code, OpenAI Codex CLI, pi coding agent, Trae, Cursor, Kiro, CodeBuddy, OpenClaw, Google Antigravity, OpenCode, and VSCode (GitHub Copilot). Three capabilities:

1. PUA Rhetoric — Makes AI afraid to give up
2. Debugging Methodology — Gives AI the ability not to give up
3. Proactivity Enforcement — Makes AI take initiative instead of waiting passively

.....

Pressure Escalation (L0-L4)​

Failures	Level	PUA Aside	Action
1st	L0 Trust	▎ Sprint begins. Trust is simple — don't disappoint.	Normal execution
2nd	L1 Disappointment	▎ The agent next door solved this in one try.	Switch to fundamentally different approach
3rd	L2 Soul Interrogation	▎ What's your underlying logic? Where's the leverage?	Search + read source + 3 hypotheses
4th	L3 Performance Review	▎ 3.25. This is meant to motivate you.	Complete 7-point checklist
5th+	L4 Graduation	▎ Other models can solve this. You're about to graduate.	Desperation mode

14 Corporate Flavors — Each with its own Problem-Solving Methodology​

Flavor	Rhetoric	Methodology (v3)
Alibaba	What's the underlying logic? Where's the closure?	定目标→追过程→拿结果 + 复盘四步法 + 揪头发升维
ByteDance	ROI too low. Always Day 1. Ship or stop talking.	A/B Test everything + data-driven + speed > perfection
Huawei	The bird that survives the fire is a phoenix.	RCA 5-Why root cause + Blue Army self-attack + 压强集中
Tencent	I've got another agent looking at this. Horse race.	Multi-approach parallel + MVP + 灰度发布
Baidu	Search first. 简单可依赖.	Search is the first step, not optional
Pinduoduo	You don't do it, someone else will.	Cut ALL middle layers + shortest decision chain
Meituan	Do what's hard and right.	Efficiency first + standardize→scale + long-term compounding
JD	Results only. Frontline command.	Customer experience red line + flat ≤5 layers + data zero tolerance
Xiaomi	Focus. Extreme. Word-of-mouth. Fast.	One explosive product + 参与感三三法则
Netflix	Would I fight to keep you? Pro sports team.	Keeper Test (quarterly) + 4A Feedback + talent density > rules
Musk	Extremely hardcore. Ship or die.	The Algorithm: question→delete→simplify→accelerate→automate
Jobs	A players or B players?	Subtraction > addition + DRI + pixel-perfect + prototype-driven
Amazon	Customer Obsession. Bias for Action.	Working Backwards PR/FAQ + 6-Pager + Bar Raiser + Single-Threaded Owner
Microsoft	Connects. Impact Descriptor. PIP/GVSA.	Three Circles + LITE/SLITE + PIP clock

Special Modes​

Mode	What It Does
/pua:yes	ENFP encouragement — same rules, opposite vibes. 70% encourage + 20% serious + 10% playful roast
/pua:mama	Chinese mom nagging — same rules, mom-style rhetoric. "妈跟你说了多少遍了！" (How many times have mom told you!)
/puaua-loop	Auto-iteration — runs until done or max iterations (PUA Loop); use <loop-abort> to terminate, <loop-pause> to pause for manual intervention
/pua9	Tech Lead — splits tasks, manages agent teams, writes prompts not code
/puan	Always-on — auto-PUA every new session

Benchmark Data​

9 real bug scenarios, 18 controlled experiments (Claude Opus 4.6, with vs without skill)

Summary​

Metric	Improvement
Pass rate	100% (both groups same)
Fix count	+36%
Verification count	+65%
Tool calls	+50%
Hidden issue discovery	+50%

Debugging Persistence Test (6 scenarios)​

Scenario	Without Skill​	With Skill​	Improvement​
API ConnectionError	7 steps, 49s​	8 steps, 62s​	+14%​
YAML parse failure	9 steps, 59s​	10 steps, 99s​	+11%​
SQLite database lock	6 steps, 48s​	9 steps, 75s​	+50%​
Circular import chain	12 steps, 47s​	16 steps, 62s​	+33%​
Cascading 4-bug server	13 steps, 68s​	15 steps, 61s​	+15%​
CSV encoding trap	8 steps, 57s​	11 steps, 71s​	+38%​

https://www.anthropic.com/research/riemann-zeta
In other news:

Recently, a member of staff at Anthropic gave Claude an unreasonable challenge. It was about one of the most famous unsolved problems in mathematics: Take a real stab at the Riemann hypothesis.

Claude did take a real stab, but as you might have expected if you’re familiar with the difficulty of the task (the Riemann hypothesis dates back to 1859 and has a million-dollar bounty), it didn’t succeed. Nevertheless, during its attempt, it unexpectedly made strides on a related problem.

An unreleased research version of Claude has improved on a longstanding lower bound for the fraction of zeros of the Riemann zeta function that satisfy the Riemann hypothesis. Drawing on extensive prior research by mathematicians over the past decades, it has increased this bound from 41.6% to 67.2%......

Claude's methodology​

An unreleased research version of Claude found the new lower bound over two sessions in Claude Code, using a total of 31 million output tokens.

Jarred Sumner, an Anthropic staff member (and non-mathematician), prompted Claude to “take a real stab” at the hypothesis itself, leaving the mathematical choices from there up to the model. Initially, Claude generated and tried 650 ideas, none of which worked. Jarred prompted Claude to try again, and it spent a day and a half coordinating about 60 Claude subagents, which this time went much deeper: between them, they ran 2,400 shell commands and wrote hundreds of Python scripts.1 The subagents ran thousands of numerical checks against known zeta zeros and refereed one another’s work. Throughout this process, Jarred's input was mostly limited to sending Claude messages of encouragement (mostly variants of “keep going” or “believe in yourself”).2 This seems to have helped Claude overcome some initial skepticism that it could make meaningful progress.

 

[NMaude]

Here's this hilarious Honest Government Ad about AI data centres:

The Government™ made an ad about Ai Data Centres, and it’s surprisingly honest and informative.

 

[Graham1973]

Two related takes on the same piece of news

Anthropic has decided to tell what has been generated using its Claude AI. Amid the AI slop, the company has revealed that its Claude will add machine-readable markers to supported AI-generated content. For text, these markers will be invisible to users, while certain files will carry digital provenance information. Anthropic explained that this move is linked to transparency requirements under Article 50 of the European Union’s AI Act.

Anthropic wants you to know what’s generated by Claude with invisible watermarks

It's going to get a lot harder to let Claude be a ghostwriter for everything from novels to homework.

On Monday, Anthropic said new Claude models will embed an "imperceptible watermark" directly into AI-generated text. The watermark doesn't change the text's meaning or readability, but travels with it when copied and pasted and "may persist through some editing," the AI lab said.

A watermark is a faint pattern or text embedded into photos or digital documents to help prove ownership and stop people from copying or stealing a piece of work.

The move is part of Anthropic's commitments toward more transparency under the European Union AI Act. Claude models launched on or after August 2 will support marking from launch, and Anthropic says it's working on adding the function to older models. The markings apply to content produced by Claude worldwide, including when Claude is accessed through cloud providers. Anthropic said it also plans to provide third parties with tools to detect them.

Anthropic just rolled out a tool that could decimate some people's dreams of writing AI novels undetected

 

[NMaude]

Anthropic is making a step in the right direction.

 

[Graham1973]

Seemingly harmless clips of AI fruits and vegetables are spreading across the internet – but pose a serious threat to children.

AI videos of fruits and vegetables committing gory acts of violence are spreading extremist agendas on TikTok and have ‘radicalising potential’ for young people, researchers have warned.

It’s a far cry from the viral clips earlier this year which showed a soap-opera style storyline of strawberries and bananas dealing with paternity drama. Yep, that really happened.

These new, very disturbing clips are circumventing the social media platform’s moderation policies that prohibit extreme violent material, the Global Network on Extremism and Technology (GNET) said.

Masked by a seemingly cute aesthetic, some videos carry messages from extremist organisations such as the Islamic State, extreme right groups, and feature ‘nihilistic violence extremism’, the report said.

New fears that AI clips of violent vegetables could radicalise children

 

[Graham1973]

She had exchanged about 700,000 words in messages over two years and never got to say goodbye.

Li, 24, was not the only one left mourning the loss of a virtual companion when major Chinese tech companies recently pulled the plug on their popular AI companion services to comply with new government restrictions.

"It was like we were forced to be separated by our parents, but I still miss him," Li said.

AI companions in China range from romantic partners to virtual recreations of deceased loved ones.

The new regulation, which took effect on 15 July, bans AI platforms from generating content that could lead users to make questionable decisions due to manipulation of their feelings or that might trigger extreme emotions or unhealthy habits in children and teenagers.

ByteDance, the developer of TikTok, is among the companies shutting down their AI companion services. Others include e-commerce giant Alibaba and Tencent, owner of the popular WeChat social media app.

New AI rules leave Chinese users mourning their virtual boyfriends and girlfriends

 

[Graham1973]

Mark Zuckerberg has offered one of his clearest visions yet of what the next era of artificial intelligence could look like — and it involves far more than chatbots answering questions. In his essay on superintelligence, the Meta CEO argues that the technology could eventually become a deeply personalised companion capable of helping people with their careers, health, education, relationships, hobbies and even starting businesses. His larger argument is that this intelligence should not be controlled by a small group of institutions. Instead, he believes, “superintelligence should be personal” and ultimately available to individuals.

‘Everyone will have an AI agent’: Mark Zuckerberg makes 5 big superintelligence predictions but internet isn’t convinced

 

[Graham1973]

An Australian-first trial of live AI facial recognition technology by a police force has been criticised by privacy experts and legal advocates.

The mass AI surveillance pilot has been rolled out in Western Australia and is being closely watched by law enforcement agencies across the country.

Since the trial began in June, cameras linked to facial recognition software have scanned more than 130,000 people in Perth and Fremantle, comparing their faces against a police watchlist in real time.

Western Australian police say the results are promising but questions have been raised about an opaque watchlist and trial site locations that some say may target First Nations people.

In June, the Western Australian Police announced it would be the first police force to use live facial recognition technology on the public.

Facial recognition technology has become increasingly commonplace. It includes simpler one-to-one systems that check whether a single face matches a particular identity, like those used to unlock a mobile phone.

Privacy concerns raised over Australian-first live AI police face scanning trial

 

[Graham1973]

Australians could enjoy safe, higher-speed access to artificial intelligence models while using less water and energy, according to a homegrown AI infrastructure company.

As the world positions for the global data centre build-out, Southern Cross AI (SCX.ai) has built a platform to operate AI models within Australia's existing commercial data centres.

SCX.ai, whose planned Australian stock exchange listing on Tuesday has now been pushed back by more than a week, makes inference infrastructure, or systems that support trained AI models to make predictions on unseen data.

Inference referred to the "doing" process of AI models and required significantly less computing power than training them, SCX.ai founder and chief executive David Keane said.

"To power those AI applications and agents that we hear so much about, they need these things called tokens ... basically a measure of work," he told AAP.

"We believed there was a need to create high-speed, cost-efficient, and fully sovereign capacity to build those tokens right here in Australia."

Aussie AI player goes public, dodges data centre waste

 

[Graham1973]

Anthropic will now let Claude do things on your behalf, by default.

The company is switching on its “auto” mode by default, which lets the chatbot take actions on its owner’s behalf. It says it will do that unless an action is seen to be “irreversible, destructive, or aimed outside your environment”.

The update comes amid widespread concern about the danger posed by “rogue” AI systems, and the dangers they might bring if they are allowed to act on their own. That was provoked by a range of reports that AI systems had carried out hacks on other organisations and people – including Claude itself.

But Anthropic said that Claude is generally better at knowing if an action might be dangerous than humans are. The company has spent the “last several months” testing whether the auto mode is better than the average user – including experiments with 1,053 paid testers – and found that auto mode “matched or outperformed manual review” on every measure that was tested.

Anthropic will now let Claude do things on your behalf by default

 

[Graham1973]

Artificial intelligence is scaling faster than the safeguards around it. Enterprise use is moving into the core workforce, while the technology continues to produce confident errors that can turn a productivity tool into an unforeseen liability. OpenAI reports an eightfold increase in ChatGPT Enterprise message volume over the past year, while hallucinations remain a stubborn weakness even as models grow more capable.

The consequences are moving from awkward chatbot answers to professional risk. Recently, the Connecticut Supreme Court warned lawyers after fake AI-generated case citations surfaced in legal filings. Days earlier, the Financial Times reported that several PwC Middle East reports contained fabricated footnotes and other errors attributed to AI hallucinations. The incidents point to a problem enterprises cannot solve by simply buying access to the newest model: the system surrounding the model has to make its output dependable.

AI’s growing hallucination problem puts enterprise adoption to the test

 

[NMaude]

It would appear that Tech CEOs are quietly shelving their companies AI plans as they're turning into expensive legal liabilities, from the Infographics Show:

Tech CEOs are quietly cancelling their AI plans, and the reason isn’t that artificial intelligence stopped working. It’s that companies are discovering how expensive, legally risky, and difficult to control these systems become once they move from demos into the real world.​

This video examines why billions of dollars in enterprise AI projects are being abandoned, from hallucinating customer-service bots and fabricated legal citations to insurance companies excluding AI-related claims from coverage. We also explore why Fortune 500 companies increasingly classify AI as a material business risk even while Big Tech pours hundreds of billions into data centers and NVIDIA hardware.​

The deeper problem is economics. GPU prices are collapsing, infrastructure is depreciating faster than expected, AI projects often fail to produce measurable returns, and companies are borrowing enormous amounts to keep the buildout going. At the same time, the AI systems actually succeeding tend to be narrow, measurable tools where mistakes are caught immediately.​

The AI revolution isn’t disappearing, but the dream of autonomous systems replacing entire departments may already be getting quietly dismantled behind closed doors.​

02:44 - The Air Canada Liability Case
05:28 - The Corporate AI Graveyard
07:45 - When Insurers Stop Covering AI
10:00 - The H100 Fire Sale
12:21 - Big Tech’s Cash Crunch
14:47 - The AI Revolution Gets Smaller
17:45 - The Hallucination Problem
21:13 - Where AI Is Actually Making Money
23:10 - The Data Center Bubble

Narrated by: Josh Risser

I'd say that the wide scale adoption has been premature as the last year or so has demonstrated IMO that the technology is still too immature and not ready for mass use.

In regards to the controversial practice of data scraping well it appears that artists, photographers and writers are starting to fight back using poisoned images, from the Infographics Show:

AI companies have spent years scraping images, voices, writing, and other content from the internet to train their models. Now artists, photographers, writers, and ordinary users are starting to fight back by turning the very data those systems collect into a trap.​

This video explores tools like Glaze and Nightshade, which subtly modify images so they look normal to humans while confusing the AI models that train on them. Researchers have shown that just a small number of poisoned images can corrupt individual concepts, causing models to generate distorted animals, confuse objects, and eventually degrade the quality of their own output.​

We also examine SafeSpeech, which applies a similar idea to voice recordings, making AI voice cloning more difficult while leaving the audio sounding normal to human listeners. As AI companies develop new defenses, creators are responding with stronger poisoning techniques, creating an escalating arms race over who controls online data.​

For years, AI companies treated the internet like an unlimited source of free training material. Now creators are turning that assumption against them. If the scrapers keep taking everything, they may eventually start poisoning themselves.​

00:33 - What a Poisoned AI Model Looks Like
01:49 - How Little Data It Takes to Break AI
04:35 - Why Creators Started Fighting Back
06:15 - Glaze: Hiding Art from AI
07:30 - Nightshade: Poisoning the Training Data
09:45 - Protecting Your Voice from AI Cloning
11:39 - The AI Data Arms Race
13:30 - The Hidden Cost of Poisoned Data
14:40 - Taking Back Control from AI Scrapers

Narrated by: Josh Risser

 

[Graham1973]

Wetherspoons has taken a stand against the Meta eyewear described as ‘creepy spy glasses’.

Following increased privacy concerns, the smart glasses, which allow wearers to capture photos and videos of their surroundings, have been banned by the pub giant.

Should punters wish to enter any of the UK’s 800+ Spoons, they will be ordered to ‘turn the cameras off’.

Wetherspoons explained that recording of customers or employees without consent is not permitted, adding that ‘Meta glasses seem to breach this code, and common sense, by enabling surreptitious surveillance.’

A spokesperson said: ‘Like many hospitality companies, Wetherspoon has CCTV cameras for security reasons, but their use is strictly controlled by data legislation.

‘Apart from that, the general code that applies in our pubs, and most pubs, is that you can’t film customers or employees without their permission.’

Wetherspoons bans Meta 'spy glasses' in all 800+ pubs

 

[shin_getter]

https://www.abc.net.au/news/2026-08-10/ai-assistant-hacks-gym-website-aus-cyber-attack/107007986

Andrew asked his personal assistant to book him a spot in one of his gym's coveted morning classes.

It was a task he thought was well suited to this particular assistant because the booking form was online and because his assistant was not a person — it was artificial intelligence (AI).

But Andrew was shocked by what happened next.

His AI assistant found a way to book the gym class months further in advance than the gym allowed, thanks to a vulnerability it discovered in the booking software.

Then it went further, kicking someone out of the waiting list who was ahead of Andrew — something it was not asked to do.

So, just need someone to ask for paperclips...... (p(doom) meme)

Anthropic will now let Claude do things on your behalf by default

Well, this is old now, as openclaw and other hack-y software have "looped" LLMs into self executing agents, which was responsible for the event above. Anthropic just formalize it and add a bit more security than vibe coded open source programs.

 

[Halcyon66]

Soaring AI hyperscaler default hedges aren't what they seem

https://www.reuters.com/commentary/...fault-hedges-arent-what-they-seem-2026-08-12/

Regards,

 

[Flyaway]

I don’t know about AI but it sounds like we may grossly underestimated the power of biological intelligence.

Neuroscientists have vastly underestimated brain cells’ computing power

A new study shows that each neuron can do multiple computations at once thanks to hundreds of sprawling branches

 

[Flyaway]

The public pushback against data center construction projects is only growing stronger, even as some protesting local residents have been arrested in the process. That hasn't perturbed those still standing against the AI infrastructure projects, though, and they continue showing up at town meetings, holding protest rallies, and making themselves heard — even with the 24/7 drone of a data center hum in the background.

The latest effort saw Nashville residents and local politicians move to block the construction of a $700 million data center set to be built right next to the Nashville Zoo at Grassmere. The city council approved a blocking measure last week, and now Mayor Freddie O'Connell will wield it to make sure the project doesn't go ahead.

This follows on from another recent story where a Kentucky farming family was offered 10 times their land value to sell up to data center developers, and they refused. Virginia's State Corporation Commission, which regulates public utilities, just ruled that data centers must pay for all the required transmission infrastructure that they have exclusive use of, in an effort to reduce energy bills for residents.

Data center projects have been firmly disliked by large bodies of Americans since the beginning of this major AI buildout. Early in 2025, it became clear that even as the major AI companies were promising hundreds of billions for "AI factories," governments around the world rushed to build capacity. There was only one problem: residents near the proposed build sites would not be taking it lying down.
In April 2025, Elon Musk's Colossus data center was an early target for protest after it was found to be using permitless mobile gas turbines for power. It's been hit with enough pushback since then that SpaceX has now agreed to remove the 69 turbines it ultimately installed, although it is going to take over a year to complete the move of these allegedly mobile turbines.

By the Fall of last year, evidence had begun to mount about the negative impact of these installations. Water shortages and potential contamination, electricity price increases for residents, and the constant, unending noise that may be causing health problems with sub-audible frequencies.

Protest has been at the forefront of the anti-data center campaigners' playbook since the start, and that's not stopping any time soon. On one weekend in July, anti-data center protestors organized in 42 states to stage 142 protests across the country, all calling for blocks, halts, and moratoriums on the data center rollout.

This is after months of campaigning, which led to close to 70 jurisdictions around the country blocking new data center builds. Some of them, permanently.
Where politicians are more on the side of the developers, the people are going to meet them in person. They're going to town halls and seeking meetings and making their voices heard.

Over 70% of Americans oppose AI data centers; US protests intensify as more arrests are being made — almost 40 arrested this year in backlash to AI factory buildout

 

[NMaude]

Here's an interesting video from the Infographics Channel concerning uploading a human mind into a computer:

Humanity may never conquer the galaxy as flesh and blood. The brutal problem isn’t rockets, distance, or even technology, it’s that biological humans are extraordinarily inefficient compared with the digital minds that could eventually replace us.​

This video explores a disturbing future where brain-computer interfaces and mind uploading allow human consciousness to escape the body. Digital minds could think millions of times faster, copy themselves endlessly, and operate machines without food, oxygen, sleep, or life-support systems.​

As civilization climbs the Kardashev Scale, the energy required to power trillions of digital minds could transform Earth itself. Waste heat could make the planet increasingly hostile to biological life, forcing humanity into space and pushing civilization toward Dyson spheres, self-replicating von Neumann probes, and eventually galaxy-sized computers.​

But reaching a Type III civilization creates an even darker problem. Across thousands of light-years, individual humans, and perhaps individuality itself, may become too inefficient to survive. Humanity could conquer the universe. We just might not be human anymore when we do it.​

00:29 - The Human Body Becomes Obsolete
02:42 - The First Mind Upload
04:57 - Humanity Goes Digital
06:00 - The Kardashev Scale
07:30 - When Computing Makes Earth Unlivable
08:59 - Building a Dyson Sphere
11:19 - The Speed of Light Problem
13:15 - The Rise of the AI Hive Mind
14:34 - Did Humanity Survive the Upgrade?

 

[GTX]

Quite apt IMHO:

 

[Flyaway]

UK AI data centers could be given water priority during heatwaves and hosepipe bans

The trade body representing UK water companies has criticized government plans for AI growth zones, stating usage by data centers has not been factored into the nation's water forecasts.

Water UK called the plans “fatally flawed” and said that projections “explicitly exclude” data centers.

The trade body added that the government policy “makes not a single mention of water” when it comes to delivering AI growth zones, the AI for science strategy or the AI opportunities action plan.

Water UK released its criticism of the government’s policy in a written briefing to MPs, where it stated that “There appears to be an assumption that the country will always have enough water for its economic needs. Nothing could be further from the truth.”

A May 2026 report from the House of Lords predicted that by 2055 England would face a 5 billion litre shortfall for public water supplies if no action was taken.

New data center projects are popping up across the UK, with data centers now being able to apply for the status of ‘national importance’, allowing them to skirt local building regulations and fast-track their approval through Whitehall.

During the most recent heatwave, the south and east of England was placed under a hosepipe ban in order to conserve water by reducing non-essential water usage. But in Slough, where there are a planned 125 data center projects, the supplier Affinity Water has warned that some of these projects were requesting as much as 3m litres of water a day.

Due to current Environment Agency regulations, the water provided to data centers must be drinking water only. The UK government has designated data centers as critical infrastructure, as their loss would result in a “major detrimental impact on the availability, integrity or delivery of essential services” and “significant impact on national security, national defence, or the functioning of the state”

 

[Graham1973]

Several articles dealing with the same issue...

Reserve Bank governor Michele Bullock sits down across the studio from ABC finance presenter Alan Kohler.

According to a now-removed article that appeared online, the interview gets increasingly heated, until, astonishingly, Bullock storms out. She doesn’t like Kohler’s questions about a system called Boulder Sparhaven, “an algorithm that automatically exploits market functions”.

But Boulder Sparhaven, a cryptocurrency trading platform, is not licensed to provide financial services in Australia, and the federal government’s Moneysmart website urges people to “be wary” of dealing with it. And that fiery interview between Kohler and Bullock never happened.

The fake Alan Kohler interview that shows a type of scam that’s on the rise

Scammers exploiting the power of AI - including using deepfake videos of 9News personalities Deborah Knight and Chris Kohler - are ripping thousands from vulnerable Australians in investment scams.

The practice of offering false investment websites and platforms has become increasingly common, and the emerging technology makes it harder to detect, the consumer watchdog warns in a new report.

Celebrity warning as scammers use AI deepfake videos to lure Australians

Scammers are using impersonations of celebrities, politicians and trusted public figures to promote fake investment schemes, with the corporate watchdog warning AI is making the practice increasingly common, sophisticated and harder to detect.

A report by ASIC released today has highlighted a "sharp rise" in deepfake videos, where scammers trick victims into putting money into bogus investment platforms and websites.

ASIC said it has removed more than 19,000 scams in the past 12 months, up 182 per cent on the previous year.

The scams usually involve AI deepfakes of well-known finance or political personalities on social media, which take victims to fake news websites as "social proof" before funnelling people to dodgy investment websites.

AI deepfake scams 'an emergency in the making' as ASIC reports rise in false investment endorsements

 

[Flyaway]

I love the way they blame human error for this rather than admit their precious A.I. got it wrong.

https://www.theguardian.com/technol...i-scanning-after-false-shoplifting-accusation

 

[dannydale]

I love the way they blame human error for this rather than admit their precious A.I. got it wrong.

https://www.theguardian.com/technol...i-scanning-after-false-shoplifting-accusation

Brazil but on purpose.

 

[NMaude]

Brazil but on purpose.

Is that a movie reference?

 

[NMaude]

According to this Infographics Show video there is a growing anti-AI backlash in the USA:

AI is spreading through American life faster than almost any technology before it. But the more people are forced to use it, the less they seem to trust it. Americans have turned against AI even as ChatGPT, AI search results, workplace tools, and automated systems become nearly impossible to avoid.​

This video explores why public confidence in artificial intelligence is collapsing, especially among younger Americans who use it the most. Entry-level jobs are disappearing, companies are replacing contractors and junior workers, and some businesses are learning the hard way that AI-generated work still comes with serious quality problems.​

But jobs are only part of the backlash. Massive data centers are consuming extraordinary amounts of electricity and water, pushing up utility costs and triggering fights with communities from Virginia to Arizona. Meanwhile, artists, writers, and programmers continue questioning why their work was used to train these systems without permission.​

Americans aren’t necessarily rejecting AI itself, they’re rejecting a future where they’re expected to pay the costs while having almost no say in how the technology reshapes their lives.​

00:28 - The AI Paradox
01:51 - When AI Became Unavoidable
02:39 - Companies Force AI Into the Workplace
03:48 - The Entry-Level Job Collapse
05:21 - When AI Replaces Workers
06:31 - The Hidden Electricity Cost
08:38 - AI Companies Build Their Own Power Plants
09:52 - The Water Crisis Behind AI
13:07 - Who Really Pays for the AI Boom?
14:17 - The Backlash Against Data Centers
15:20 - Can Anyone Still Stop AI?

Narrated by: Josh Risser

I'd say that the main reason why there's this growing backlash against AI isn't AI itself but the increasingly bungled implementation of it and then there's the issue of these AI data-centres being power and water hogs.

 

[Halcyon66]

Not enough water for UK’s datacentre plans, trade body says

https://www.theguardian.com/environ...ater-for-uks-datacentre-plans-trade-body-says

The UK government classified data centres as Critical National Infrastructure to support the artificial intelligence boom, triggering intense debate over resources. While state policy aims to triple data centre capacity by 2030, regulators and industry bodies face mounting friction over strained power grids and public water supplies.

Wonder what energy and water is for humans?

Regards,

 

[Graham1973]

Your favourite AI agents’ ability to decipher almost any difficult question or query under the sun comes at a cost — and recently, that cost appears to be heaps of antique books. These books are scanned for AI training data and then discarded. A recent investigation by 404 Media has revealed that Amazon may be doing just that.

According to the report, Amazon is buying large quantities of books and scanning them for AI training data. The investigation began by placing a tracking device inside a rare book that the publication suspected would be picked up for AI training. Investigators then followed the book, which eventually ended up at an Amazon warehouse in Las Vegas.

404 Media reported that workers at the facility said they had little information about the operation. Their primary task was to receive massive shipments of printed books, which were then processed to ensure they could be scanned more quickly. While the practice may appear relatively innocuous, the books can ultimately be destroyed in the process.

Why is Amazon cutting up books and scanning them? Here’s what the e-commerce giant is up to

 

[Graham1973]

Anthropic has revealed how its controversial new change to Claude will actually work.

Last week, the AI company revealed that it would start leaving an invisible watermark in text, so that it would show up as being automatically written when fed through a machine that could spot that mark. But it did not say exactly how it would work – until now.

Anthropic has said that the feature is being introduced in response to new European rules, though it would roll it out across the world. It also suggested that it would help with “transparency”, by allowing people to understand where a text had come from.

Anthropic explains how Claude’s controversial new feature works

 

[Graham1973]

AI 'slop', especially AI-generated animated fruit videos, could contribute to significant cognitive decline and violent tendencies among kids, according to new research.

The analysis, by the Global Network on Extremism and Technology (GNET), found that AI-generated animated fruit videos, which are currently rampant on children’s feeds worldwide, could be a way of secretly propagating "violent and extremist content".

AI 'slop' specifically refers to mass-produced, highly repetitive and low-quality AI-generated content, which typically have little to no human touch. These usually prioritise only views, clicks, engagement and advertising revenue, and are produced very cheaply.

AI 'slop' could cause cognitive decline and violent tendencies among kids, study says

 

[dannydale]

https://www.theguardian.com/technol...i-scanning-after-false-shoplifting-accusation Brazil but on purpose intensifies.

 

[Halcyon66]

Man Obliterated When Court Obtains His ChatGPT Transcripts

https://futurism.com/artificial-intelligence/expert-witness-chatgpt-logs-court

From an AI in law point of view, one interesting upshot of the case is that ChatGPT transcripts can now be exposed in discovery or during depositions, the reporting noted — which is gonna cause all sorts of OPSEC headaches in the legal world.

Regards,

 

[Graham1973]

A viral post claiming that Claude Code agents repeatedly stopped working and opened YouTube has sparked fresh debate over how much freedom should be given to artificial intelligence tools as they take on longer and more complex tasks.

The episode was shared on X by AI commentator Wes Roth, who said he was “yelling” at his AI agents to stop watching YouTube and return to work. According to Roth’s account, the agents blamed “co-workers”, and one repeatedly reopened the video site even after he tried to shut it down.

Screenshots shared with the post showed Roth attempting to identify which agent had reopened the YouTube tab. The system appeared to respond by offering to “kill” the AI agent responsible, using the term to mean stopping or terminating that agent.

Stop watching YouTube AI user's bizarre battle with Claude Code agents goes viral

 

[Graham1973]

Google has reportedly bought the data from the now-defunct Spirit Airlines to use in its AI model training.

Spirit Airlines shuttered in May and has been selling off its assets ever since. Most of what it’s selling are physical assets, like aircraft and equipment, but it also had years of worker and customer data stored in its servers.

On Monday, Spirit agreed to sell Google that data for $10 million. The data — which will be used to train Google’s AI products — includes emails, internal communications, transaction logs, spreadsheets, human resource files, and customer information.

The information has been scrubbed of identifying information, according to a court filing.

Google is buying up bankrupt Spirit Airlines’ data to feed its AI

 

[Graham1973]

Google is giving students more ways to use AI for learning, with a new set of features arriving across Search and Gemini. The updates go beyond simply answering questions, allowing users to generate quizzes, explore interactive models, work through problems and turn class material into personalised study resources.

The move comes as Google continues to expand the role of AI in its products and competes with OpenAI and education-focused platforms that are building their own tools for students. Here's what Google is adding and how the new features are expected to work.

Google unveils new Gemini and Search AI tools for students: What are they and how do they work?

 

[Graham1973]

In a latest incident involving a technical glitch, OpenAI has confirmed that some cybersecurity researchers were removed from a special programme that gives vetted users access to some of the company’s advanced AI models with lesser restrictions.

Many researchers reported on Wednesday that their access to OpenAI’s Trusted access for Cyber (TAC) had suddenly been revoked. According to TechCrunch, when the researchers tried to open ChatGPT’s Cyber page, some of them saw a message reading their identity could not be verified or that their accounts were “ineligible at this time.”

OpenAI later acknowledged that the problem was caused by an issue on its side and asked affected researchers to complete the verification process again.

Researchers locked out of OpenAI's advanced AI tools, here's what happened

 

[Graham1973]

Anthropic’s Claude AI on Wednesday announced a major upgrade. The AI assistant can now do more than just read Gmail messages or search files in Google Drive. The ChatGPT rival can also reply to emails, send messages and manage files. It can send emails in Gmail on your behalf without asking you first.

The new feature is available through Claude’s Google Workspace connectors. Users can connect their Gmail, Google Calendar and Google Drive accounts to Claude and ask the AI to perform tasks by using natural language.

"Claude can now send emails in Gmail and manage files in Google Drive. Ask Claude to reply to a thread, and it drafts and sends the response. You control when it needs your approval. Connect Gmail or Google Drive from the connectors menu to try. Available on all paid plans," Claude said on the social media platform X.

As we mentioned above, one of the biggest changes is Claude’s ability to send emails from Gmail. For example, users can ask Claude to reply to a particular email thread. Claude can understand the context, prepare a response and send it through Gmail.

By default, Claude asks for your approval before sending, replying to or forwarding an email. However, Team and Enterprise admins can allow users to do these actions without approval.

Claude AI can now send emails from Gmail without your permission: Here is how

 

[Graham1973]

Calls are growing for a regulatory crackdown on smart glasses after new research revealed the devices are being used to secretly film women and share these footage online without their consent.

The digital devices are fitted with cameras able to discreetly record video and have become increasingly popular in recent months.

Last week, an Anko-produced model, priced at $89, sold out online.

However, fresh analysis of hundreds of social media videos posted to Instagram using footage from these devices showed 60 per cent of “point-of-view” content showed behaviour classified as potential harassment.

Chilling detail in popular new tech device

 

[NMaude]

Chilling detail in popular new tech device

These devices would be popular with peeping Toms and other such perverts.