Thanks for the link. A well written article.
1 Like
Sooner or later someone was going to look under the hood and see what was open.
One expert put it succinctly,
Julia Powles, an associate professor of Law and Technology at the University of Western Australia and the director of the Minderoo Tech and Policy Lab, says programmers have to train AI technology such as ChatGPT to behave ethically.
“These are not reasoning machines, they’re word-prediction machines,” Dr Powles explains.
“Because they have no concept of what the words they generate mean, they simply have no capacity to reason ethically.”
“Upgrade” at your peril!
‘“The ultimate upgrade” – Doctor Who & the Cybermen (parts 1 & 2) | BioethicsBytes
2 Likes
Bring on the Butlerian Jihad.
Fun with AI. Trust me! 
A wonderful example of the world’s growing industrial might. Remebering that in every contested case, one or other of the lawyers stuffs up and loses, A.I. simply provides new help in this process. As happens all the time, lawyers present the wrong law, or the wrong cases; A.I. now allows them to do so on an industrial scale.
How long will it be before the AI bots displace humans? The silliness of many internet articles already shows how inane they can be, but Microsoft might get the gong. How did I miss this?
edit: Another spin to the story
1 Like
As with a lot of technology AI is being sold on its smoke more than its smell. This is not the only recent report about blind faith in what is at the end of the day a very fancy search machine coupled to linguistic translators.
There is promise of AI but the terminology is being applied to ages old rules based processes as well as gaming methodologies, not just the emergent futures of protagonists hope and expectations it can and will do, save for refusing to ‘open the door’. It is not intelligence nor sometimes more accurate than a first week staffer having had an hours training.
There is more to this case than the present limits of AI. How did such accusations get published without a human checking them? What kind of review mechanisms do they have that allow this to happen?
In another 30 years we will have a better idea what areas AI is useful in and how to verify its material. There is going to be much heartache if AI is found to be unable to meet reasonable standards in some areas where the marketeers so desperately want it to and where they already assume it will work.
In future I can imagine self-guided vehicles could have big disclaimers that in some environments all guarantees and promises are voided, you have no insurance and you are on your own.
We might not be fully aware of how AI is being used by our GP in a clinical practice. Would save time by writing notes, referrals, analysing scans and interpreting data to aid in diagnosis. Apparently very much used in the European countries, so much so that legislation is being called to ‘cover’ health professionals.
The following is an interesting article on the risks/dangers of relying on AI. Don't rely on Dr AI: New search engine gives medical advice that could lead to death in one in five cases | Daily Mail Online
It comes as medics last month were warned they could be risking patient safety by relying on AI to help with diagnoses.
Researchers sent the survey to a thousand GPs using the largest professional network for UK doctors currently registered with the General Medical Council.
One in five admitted using programmes such as ChatGPT and Bing AI during clinical practice, despite no official guidance on how to work with them.
Experts warned problems such as ‘algorithm biases’ could lead to misdiagnoses and that patient data could also be in danger of being compromised.
They said doctors must be made aware of the risks and called for legislation to cover their use in healthcare settings.
2 Likes
Calling it AI really is a misnomer. The AI bots are trained on real world articles, graphics, photos, comments, posts etc… They are not actually being intelligent, simply learning existing information. Noam Chomsky wrote an article worth reading…
2 Likes
Chomsky uses many words to say garbage in, garbage out. This is the primary stupid loop.
Articles on the topic point out that the AI teachers, groomers, custodians go to some trouble to cull the learning material so that AI engines don’t learn from nonsense. Nonetheless some models have been shut down because they learned the wrong things. Given the vast bulk of data required for training, the process of input curation can only be superficial. Extracting the juice out of vast amounts of data is one of the purposes of AI. Requiring humans to do it for them while they are trained is the second stupid loop.
Then we have the delightful possibility that as more internet material is AI generated future generations of AI will be trained on it. This stupid loop is self-reinforcing.
Until a quite different model is created the strengths of AI will be writing perfect but vacuous prose and generating amusing images and doggerel. Oh, and stone-walling customers with complaints so saving on hiring staff to provide service.
There once was a bot in the chat,
Who talked crap, leaving customers flat.
In loops it did spin, nonsense its sin,
Leaving customers thinking, “What’s up with that?”
1 Like
Yes. I might be wrong but I don’t think that at the moment a ‘thinking machine’ has been created to ‘spell the end of the human race’ as Stephen Hawking warned.
AI doesn’t function like the human mind which looks for explanations and possibilities even from limited informations.
I can see AI programs such as ChatGPT etc as an aid to the medical profession because of the ability to collect a lot of information very quickly, but as the latest research found can also lead to dangerous conclusions if relied on exclusively.
2 Likes
He also points out that whilst we humans understand our various languages and all of its nuances… AI can never get there. We are born with the capacity to learn a language, its innate
2 Likes
