"artificial intelligence" (AI)* thread 🧵
-
Sal Khan of the Khan Academy said we will soon all have private "AI" tutors or some such headline grabbing nonsense. (Proving only, perhaps, that he and other rich people are just as stupid as the rest of us.)
Khan Academy then rolled out an "AI" tutor bot. Unfortunately it can't do math. Not complicated math. Math, as in, it doesn't know how to round. Oh good.
(That's not what these chat bots are good at.)
By the way the obvious application to the job search "AI" is to actually employ it IN ORDER TO BE racist and sexist and then blame the racism on the AI when the company is caught.
Well, or not care about the racism. (And still blame the AI.)
-
By the way the obvious application to the job search "AI" is to actually employ it IN ORDER TO BE racist and sexist and then blame the racism on the AI when the company is caught.
Well, or not care about the racism. (And still blame the AI.)
Fake quotes generated by an LLM ...
... attributed to real people.
The coming years are gonna be hella fun.
https://dair-community.social/@emilymbender/111969995986122092
-
Fake quotes generated by an LLM ...
... attributed to real people.
The coming years are gonna be hella fun.
https://dair-community.social/@emilymbender/111969995986122092
By the way Sal Khan thinks the Koch brothers are doing good work out in the world.
https://mastodon.social/@jmcrookston/111607351563409147
It's truly two worlds. The rich, and the rest of us.
-
By the way Sal Khan thinks the Koch brothers are doing good work out in the world.
https://mastodon.social/@jmcrookston/111607351563409147
It's truly two worlds. The rich, and the rest of us.
About using a chatbot as a professor, someone at Harvard said this:
' "We'll make clear to students that they should always think critically when taking in information as input," Malan told the Crimson, "be it from humans or software." '
Sounds like a plan.
Humans always think critically.
Nothing could go wrong.
️ -
About using a chatbot as a professor, someone at Harvard said this:
' "We'll make clear to students that they should always think critically when taking in information as input," Malan told the Crimson, "be it from humans or software." '
Sounds like a plan.
Humans always think critically.
Nothing could go wrong.
️ I've been seeing ads for Grammarly for years now but it's suddenly an "AI writing tool" haha
-
I've been seeing ads for Grammarly for years now but it's suddenly an "AI writing tool" haha
"Large language models (LLMs) are more likely to criminalise users that use African American English, the results of a new Cornell University study show. "
Article talks about "covert racism" - it's not covert. If you train LLMs on racist data it outputs racist outputs COME ON.
-
"Large language models (LLMs) are more likely to criminalise users that use African American English, the results of a new Cornell University study show. "
Article talks about "covert racism" - it's not covert. If you train LLMs on racist data it outputs racist outputs COME ON.
Sometimes AI authorship stands out – as in a paper from Radiology Case Reports entitled "Successful management of an Iatrogenic portal vein and hepatic artery injury in a 4-month-old female patient: A case report and literature review."
This jumbled passage is a bit of a giveaway: "In summary, the management of bilateral iatrogenic I'm very sorry, but I don't have access to real-time information or patient-specific data, as I am an AI language model."
https://www.theregister.com/2024/03/19/ai_researchers_reviewing_peers/
-
Sometimes AI authorship stands out – as in a paper from Radiology Case Reports entitled "Successful management of an Iatrogenic portal vein and hepatic artery injury in a 4-month-old female patient: A case report and literature review."
This jumbled passage is a bit of a giveaway: "In summary, the management of bilateral iatrogenic I'm very sorry, but I don't have access to real-time information or patient-specific data, as I am an AI language model."
https://www.theregister.com/2024/03/19/ai_researchers_reviewing_peers/
The hits keep comin:
Steal your employees' tips! Discriminate against renters on basis of race!
Just two of the wonderful suggestions that Microsoft's LLM is telling people to do. (Clearly simply reflecting what humans do, back to them)
https://themarkup.org/news/2024/03/29/nycs-ai-chatbot-tells-businesses-to-break-the-law
-
The hits keep comin:
Steal your employees' tips! Discriminate against renters on basis of race!
Just two of the wonderful suggestions that Microsoft's LLM is telling people to do. (Clearly simply reflecting what humans do, back to them)
https://themarkup.org/news/2024/03/29/nycs-ai-chatbot-tells-businesses-to-break-the-law
Article presented interview with Michael Schumacher and hid that it was entirely AI chatbot generated until the end. It even then suggested the AI had been trained by Mr. Schumacher.
The Schumacher family sued. The magazine had to pay $217,000.
https://www.thedrive.com/news/michael-schumachers-family-gets-settlement-after-fake-ai-interview
-
Article presented interview with Michael Schumacher and hid that it was entirely AI chatbot generated until the end. It even then suggested the AI had been trained by Mr. Schumacher.
The Schumacher family sued. The magazine had to pay $217,000.
https://www.thedrive.com/news/michael-schumachers-family-gets-settlement-after-fake-ai-interview
Google might have liability for its "artificial intelligence" snippets because it is no longer merely surfacing third party information.
https://www.vox.com/technology/351189/google-ai-overview-section-230
-
Google might have liability for its "artificial intelligence" snippets because it is no longer merely surfacing third party information.
https://www.vox.com/technology/351189/google-ai-overview-section-230
Is AI a scam? Yep.
-
Is AI a scam? Yep.
Altman recently asked people to consider a time when we'll be able to ask artificial intelligence to "discover all of physics" or "start and run a great company" and that "it can go ahead and do that."
Sigh
https://www.cbc.ca/news/business/ai-generative-business-money-revenue-1.7264014
-
Altman recently asked people to consider a time when we'll be able to ask artificial intelligence to "discover all of physics" or "start and run a great company" and that "it can go ahead and do that."
Sigh
https://www.cbc.ca/news/business/ai-generative-business-money-revenue-1.7264014
A Washington Post investigation conducted in collaboration with researchers at the University of California, Riverside, recently revealed that ChatGPT requires up to three bottles of water to generate a single 100-word email. If the global AI demand continues its current trajectory, best estimates put its water withdrawal around 4.2 – 6.6 billion cubic meters by 2027; the equivalent of the annual consumption of half the United Kingdom.
https://www.techpolicy.press/why-we-dont-know-ais-true-water-footprint/
-
A Washington Post investigation conducted in collaboration with researchers at the University of California, Riverside, recently revealed that ChatGPT requires up to three bottles of water to generate a single 100-word email. If the global AI demand continues its current trajectory, best estimates put its water withdrawal around 4.2 – 6.6 billion cubic meters by 2027; the equivalent of the annual consumption of half the United Kingdom.
https://www.techpolicy.press/why-we-dont-know-ais-true-water-footprint/
-
And here is a paper cited in the article which notes in its abstract that for Microsoft to train GPT3, it directly evaporates 700,000 litres of fresh water.
Later, they note AI consumes 0.33% of the United States' total water consumption.
-
And here is a paper cited in the article which notes in its abstract that for Microsoft to train GPT3, it directly evaporates 700,000 litres of fresh water.
Later, they note AI consumes 0.33% of the United States' total water consumption.
I stumbled on this page which explains what the various vendors use water for, how some of it is lost, how reclaimed, how water use is measured. I can't vouch for the site, I don't know it, but it seems to have very detailed information about data centre water use.
https://dgtlinfra.com/data-center-water-usage/
I should go looking for other industrial users of water to see how they compare to current state of AI data centre water use.
-
I stumbled on this page which explains what the various vendors use water for, how some of it is lost, how reclaimed, how water use is measured. I can't vouch for the site, I don't know it, but it seems to have very detailed information about data centre water use.
https://dgtlinfra.com/data-center-water-usage/
I should go looking for other industrial users of water to see how they compare to current state of AI data centre water use.
For anyone who ever wanted their social book reading app to [checks notes] aggressively accuse them of being anti-white racist in the fun end-of-year summary, well, AI has got you covered.
-
For anyone who ever wanted their social book reading app to [checks notes] aggressively accuse them of being anti-white racist in the fun end-of-year summary, well, AI has got you covered.
"AI Search Has A Citation Problem.
We Compared Eight AI Search Engines. They’re All Bad at Citing News."
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.php -
"AI Search Has A Citation Problem.
We Compared Eight AI Search Engines. They’re All Bad at Citing News."
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.phphttps://apnews.com/article/gabbard-trump-ai-amazon-intelligence-beca4c4e25581e52de5343244e995e78
"Experts had predicted the process could take many months or even years, but AI accelerated the work by scanning the documents to see if they contained any material that should remain classified, Gabbard said during her remarks at the Amazon Web Services Summit in Washington."
Didn't they accidentally release a whole bunch of Social security numbers and other sensitive data?
-
https://apnews.com/article/gabbard-trump-ai-amazon-intelligence-beca4c4e25581e52de5343244e995e78
"Experts had predicted the process could take many months or even years, but AI accelerated the work by scanning the documents to see if they contained any material that should remain classified, Gabbard said during her remarks at the Amazon Web Services Summit in Washington."
Didn't they accidentally release a whole bunch of Social security numbers and other sensitive data?
Divorce because of chat bot advice. Sure, why the hell not.
Dating a chatbot? Hell yeah we've come this far let's do this.
