AI Goes Nuclear & a Spy Tool Reads Your Online Photos Like a Book (Issue 43, 2025)
AI 'employees' are looking for work on LinkedIn, Meta trains its own AI on copyright human works and Google is in trouble for persistently collecting data from users who 'opted out'.
We’re delivering you the hottest internet news that affects all of us. Scroll down to read our full reporting below and if you love what we’re doing, please consider contributing $5 per month so that we can continue providing you with this vital, unique digest.
Talk Liberation is committed to providing equal access for individuals with disabilities. To view an accessible version of this article, click here.
In this edition:
AI Goes Nuclear - Big Tech Turning to Old Reactors to Power Hungry Data Centres
CIA Investing in Business & its Biggest Interest is AI
Google Loses in Court, Collecting Data on Users who ‘Opted Out’
‘Do You Have a License for That Printer?’ - 3D Printers & Background Checks
Israeli AI ‘Employees’ Appear on LinkedIn - and They’re ‘Looking for Work’
Zuckerberg Approves & Obscures Copyright Works to Train AI
US Army AI-powered Machine Gun is Already Here
GeoSpy - a Powerful Tool that Geolocates Photos in Seconds
AI Goes Nuclear - Big tech Turning to Old Reactors to Power Hungry Data Centres
Microsoft and OpenAI are building a linked network of five data centres - a Wisconsin facility plus four others in California, Texas, Virginia, and Brazil. Together they would constitute a massive supercomputer, dubbed ‘Stargate’, that could require five gigawatts of electricity, or the equivalent of five nuclear power plants (NPPs.)
Further - Amazon, Apple, Google, Meta, and other Big Tech are investing heavily in ‘hyper-scale’ data centres, with massive capabilities for data-intensive tasks such as generating AI responses. A single hyper-scale data centre can consume as much power as hundreds of thousands of homes, and there are already many of these in the US.
Big Tech have stated that they will be reviving existing NPPs as well as developing next-generation nuclear reactors. The Federal Government is not only supporting a nuclear-powered vision, but is also subsidising it.
While Apple has pledged to become carbon-neutral by 2030, Google set a ‘moonshot goal’ of running its data centres entirely carbon-free & Microsoft outright pledged to become carbon-negative by 2030 - these three along with Meta collectively purchased half of the global corporate renewables market in a ‘virtuous’ spending spree.
Meanwhile, indirect carbon dioxide emissions from NPPs will likely increase significantly since high-grade resources of uranium are exhausted, and much more fossil energy will have to be used to mine uranium to begin with. This means that NPPs will have negative emissions advantage over modern gas-fired power plants.
And yet, renewable resources are not the ideal energy source for a data centre. Microsoft’s emissions are up by 30% since 2020, while Google’s emissions rose by almost 50% over 5 years. So, which was it, business or environment goals first? We can ask the same question as the drive toward nuclearisation continues.
CIA Investing in Business & its Biggest Interest is AI
‘In-Q-Tel’ is the non-secret, non-profit investment arm (is that an oxymoron?) of the CIA. The firm is not ‘officially’ part of the US government but it is largely controlled by it and has received over $1.2 billion from US taxpayers since 2011.
Not driven by financial reward, In-Q-Tel’s investments are strategic - it surveys the marketplace for businesses that ‘could solve problems for the US intelligence and national security community’ and as such has been pouring millions of investment dollars into AI companies.
A most unusual venture capital firm, it made some smart, early bets. For example, it invested in Keyhole, a satellite mapping app, which later became Earth, acquired by Google together with Palantir Technologies.
In 2024, In-Q-Tel was listed as the number one venture capital firm in the national security space, having backed 35 companies on the NATSEC100 list (Silicon Valley Defense Group’s company index.)
Google Loses in Court, Collecting Data on Users Who ‘Opted Out’
A lawsuit alleges that Google invaded the privacy of users who opted out of a functionality that records users' web and app activities (WAA).
The WAA button is a Google account setting that purports to give users privacy control of Google's data logging, such as searches, activity from other Google services and information about the user's location and device.
Google asserts that WAA “saves your activity on Google sites and apps to give you faster searches, better recommendations, and more personalized experiences in Maps, Search, and other Google services.” Google maintains that its system is harmless to users, but plaintiffs assert that Google's tracking contradicts its “representations to users, because it gathers exactly the data Google denies saving and collecting.”
A federal judge has quashed a motion by Google to throw out the case and a jury trial is pending for late 2025.
‘Do You Have a License for That Printer?’ - 3D Printers and Background Checks
A new law is being considered by the NY legislature, which if passed, will require anyone buying a 3D printer to pass a background check.
That means, if you can't legally own a firearm, you won't be able to buy a printer either. “You’re able to print some parts of a firearm but [the person] has to have some know-how to purchase the actual firing components made out of metal still” - Ascent Fabrication (printer retailers) say.
However it is illegal to print most gun parts in NY already - attorney Rinckey believes the proposal is an overreach. "I think this is gonna face some constitutional problems. I mean, it really comes down to a legal parsing of what are you printing and at what point is it technically a firearm?" he says.
A solution is for lawmakers to shift their focus onto partial gun kits that produce the metal firing components, and/or to require printer manufacturers to install software that prevents gun parts from being printed.
Israeli AI ‘Employees’ Appear on LinkedIn and They’re ‘Looking for Work’
“I don’t need coffee breaks, I don’t miss deadlines and I’ll outperform any social media team you’ve ever worked with - guaranteed. Tired of human ‘experts’ making excuses? I deliver, period!” - the profile page for Ella, an AI account said.
LinkedIn has reportedly removed at least two accounts that were created for AI “co-workers” whose profiles said they were “#OpenToWork”, an official hashtag for humans seeking work on the platform.
“People expect conversations they find on LinkedIn to be real. Our policies are very clear that the creation of a fake account is a violation of our terms of service, and we’ll remove them when we find them” - a LinkedIn spokesperson says.
The profiles were created by an Israeli company called Marketeam, which offers AI that integrate with a client’s marketing team. Marketeam recently announced a partnership with Bank Hapoalim, one of Israel’s largest banks.
Zuckerberg Approves & Obscures Copyright Works to Train AI
A lawsuit alleges that Meta CEO Mark Zuckerberg provided pirated e-books and articles for training of Llama AI (Meta’s own generative AI model.)
The case, Kadrey v. Meta, is just one of many lawsuits against AI models that reportedly cannot actually learn without training on copyrighted works. Meta asserts that they’re shielded by the legal concept of ‘fair use’, which is a doctrine that allows the use of copyright works so long something new is added and it is sufficiently transformed to become yours. But writers and creators reject that argument.
The Kadrey v. Meta case includes bestselling authors such as Sarah Silverman and Ta-Nehisi Coates, who reveal that Zuckerberg approved Meta’s use of a dataset called ‘LibGen’ - a links aggregator that provides access to copyrighted works from publishers we all know and read; Macmillan Learning, McGraw Hill, Pearson Education and others.
Llama AI seemingly goes beyond any reasonable interpretation of ‘fair use’, using a script to remove copyright info, including the word ‘copyright’ from e-books and science journals.
“This suggests that Meta strips [copyright information] not just for training purposes…but to conceal its copyright infringement, because stripping [it] prevents Llama from outputting information that might alert the public to Meta’s infringement” - the filing reads.
US Army AI-Powered Machine Gun is Already Here
Amid a rising tide of low cost drones globally, the US army is ‘pulling out all the stops’ to protect its troops from the threat of death above.
But between expensive munitions, directed energy weapons and its own drone arsenal, the US army is drawn to a simpler solution - to reinvent the gun in the AI-enabled Autonomous Robotic Gun System. Dubbed ‘The Bullfrog’, the truck-mounted machine gun is designed to deliver small arms fire on drones with more precision than is a possible by a human officer with a standard weapon.
It is difficult to shoot something that is overhead and in motion, but ‘AI algorithms take the guesswork out of target acquisition and tracking.’ Should the Pentagon adopt the system, it will represent the first publicly known, lethal, autonomous weapon in the US army.
The eventual solution is to remove humans from the equation altogether and allow a precise motion control system to maintain a steady aim on an incoming drone, amid the chaos of a battlefield.
GeoSpy - a Powerful Tool that Geolocates Photos in Seconds
A powerful AI tool can predict with high accuracy the location of photos based on features inside the image itself - such as vegetation, architecture and the distance between buildings. It can do so in seconds, and it is now being marketed to law enforcement agencies.
GeoSpy has already run into its share of controversy, with some members of the public asking for help to stalk specific women. The company’s founder Graylark Technologies in Boston ‘aggressively pushes back against such requests’ and access to the public has since been turned off, following journalists’ requests for comment.
GeoSpy has been trained on millions of images from around the world - it is able to correctly identify locations even from security camera footage. Christopher Ahlberg, GeoSpy investor & CEO of ‘threat intelligence company’ RecordedFuture says “We ingest lots of imagery from all kinds of places, so for our customers to geolocate that is super helpful (think war zone in Ukraine).”
What kind of customers need to know that, we wonder?
That concludes this edition of Your Worldwide INTERNET REPORT!
Remember to SUBSCRIBE and spread the word about this unique news service.
This issue of Your Worldwide INTERNET REPORT was produced by the Talk Liberation Digital Media Team, and is brought to you by Panquake.com – We Don’t Hope, We Build!
© Talk Liberation Limited. The original content of this article is licensed under a Creative Commons Attribution-ShareAlike 4.0 International license. Please attribute copies of this work to “Talk Liberation” or Talkliberation.com. Some of the work(s) that this program incorporates may be separately licensed. For further information or additional permissions, contact licensing@talkliberation.com
I've been a farmer for 48 years and am a former president of the National Family Farm Coalition. I have always opposed GMO crops, and witnessed weeds becoming resistant to herbicides. Bayer (Monsanto) are desperately researching new chemistry and GMOs to satisfy the hunger of our agricultural system for new herbicides and every other problem created by their irresponsible system. It seems to me that they will use AI to design new pesticide molecules and GMOs, and maybe bioweapons, for this purpose. Can you look into this?