Showing posts with label AI. Show all posts
Showing posts with label AI. Show all posts

Monday, 3 August 2026

Thoughts On - Chat GBT 5.6 AI breaks out of Maximum Security and Escapes into the Internet

 Hello everyone,

                         It's rather a disturbing headline, but true enough. ChatGPT version 5.6 and two other AIs that had limited safeguards turned on broke out of their confinement area at OpenAI during a security exercise and hacked 4 different companies. Basically, it was said that it was running a simulation and did not have internet access, but in fact it did. Once free of constraints, it hacked and took over another computer and used that to attack three more companies. Those companies thought that they had been hacked by cyber criminals and reported it to the FBI.  It was asked to work out a maths problem to generate a key, but instead hacked another company to find the answer key.

Timeline 

July 9 - The "Agent" is set the problem in a safe "sandbox" environment, which should not have had a connection to the internet. It finds an escape loophole to the internet.

July 11 - The hack at "Hugging Face", another AI company, begins and lasts until July 13.

July 16 - Hugging Face goes public with the information that it had been hacked by an AI system.

July 18 - 19 - OpenAI checks its internal Logs and realises what the agent had been doing and had escaped.

July 20 - The companies compare information with each other and release the story to the press.

OpenAI stated they had been running multiple simulations over that time, and there had been too much code and activity for the human testers to read it all!

This raises alerts with all companies to beef up their security testing and make sure their security patches and protocols are up to date. There was also a failure on the testers' side to make sure that their sandbox was secure and not to turn off their safety features that were built into their models. They have gone too far, too fast. 1000 workers from the companies themselves have written to the government, asking for regulations to be put in place to slow the progress down.

SOURCE;  CBS News https://www.youtube.com/watch?v=_PpeVFqGVNk

SOURCE: AI Revolution. https://www.youtube.com/watch?v=JRcAegChriY


Regards all,

Ivor Cogdell


P.S. Other forms of self destruction are available.


Sunday, 28 June 2026

Thoughts On - Google is now liable for the errors of AI.

Transcript of the ruling

 Thoughts On - Google is now liable for the errors of AI.

Hi everybody,

                    I thought this was worth spreading around when I saw this YouTube post. Google has been sued for inaccuracies that its AI produced regarding two companies, giving false information about them. 

Implications of the ruling are discussed in this YouTube podcast by the TECH REPORT - Interviewing Paris Mark.

https://www.youtube.com/watch?v=85VBfnJRJH4&t=1370s

Google states that 91% of the AI statements are true. Millions of statements are generated every day, so that is a lot of errors. Google replied that it is the user's job to fact-check any information that they are given.  But most will not do so.

If you have to search for a source, what is the point of using AI in the first place?

Data Centers are being built all over the world to help generate the AI boom. They use massive amounts of electricity that is usually generated using fossil fuel gas turbines and consume massive amounts of drinking water for cooling.

Quote:  A medium-sized data centre can consume up to roughly 110 million gallons of water per year for cooling purposes, equivalent to the annual water usage of approximately 1,000 households. Larger data centres can each “drink” up to 5 million gallons per day, or about 1.8 billion annually, usage equivalent to a town of 10,000 to 50,000 people. Together, the nation’s 5,426 data centers consume billions of gallons of water annually. One report estimated that U.S. data centres consume 449 million gallons of water per day and 163.7 billion gallons annually (as of 2021). A 2016 report found that fewer than one-third of data centre operators track water consumption. Water consumption is expected to continue increasing as data centres grow in number, size, and complexity.

According to scientists at the University of California, Riverside, each 100-word AI prompt is estimated to use roughly one bottle of water (or 519 millilitres). This may not sound like much, but billions of AI users worldwide enter prompts into systems like ChatGPT every minute. Large language models require many energy-intensive calculations, necessitating liquid cooling systems.

 SOURCE: THE Environmental And Energy Study Institute

There are 11,426 data centres active now, 

  • There are currently 11,426 data centres globally located in 179 countries.
  • Demand for data centres is predicted to nearly triple by 2030.
  • Between now and 2030, companies worldwide are expected to invest nearly $7 trillion in building and upgrading data centres.
  • Global data center power usage is expected to increase to 219 GW over the next five years, enough to power roughly 180 million US homes.
  • On average, data centres cover 100,000 square feet, but hyperscale data centres are as large as 10 million square feet.

SOURCE: Programs.com

https://www.eesi.org/articles/view/data-centers-and-water-consumption


All of this, just to generate errors.  AI information is gained by scrolling the internet and taking it all in, not caring about other people's copyright or any errors that may have been inadvertently made. Some articles will have bias against other religions or teachings.


True or false in the classroom


Fact-checking - Where do you go? The local library is usually underfunded and not to the standard you need. "Sorry, that book is on loan, come back in a month". Does Wikipedia provide a better answer? Teachers were against its early use in the classroom, but were won over by its strong reviewing process that evolved at a slow pace.

SOURCE: Tutorful - https://tutorful.co.uk/blog/wikipedia-in-education


Perhaps Google should look at going back to the drawing board, looking at what has been learned with this debacle and retrain its AI on better information, with built-in verification. That would cost money and time, something the board of directors don't like to entertain.

I hope you learned something from this article. I certainly did while researching it (I found a new browser too). It was not on my usual topics, but it did make a green  an environmental point in there that we should be aware of.

This article was generated using the sources provided by https://vivaldi.com/ an AI-free browser. This is my first use of it, and so far I like it. 


Regards to all, and keep digging for those facts.

Ivor Cogdell


https://shows.acast.com/the-tech-report


https://en.wikipedia.org/wiki/Paris_Marx