META turned off “AI for Science” three days after launch
META deleted the public version of the Galactica AI-Algorithm after scientists convicted him to generate false and misleading information. Motherboard writes about this.
The technical gigant described the system as a language model that can “store, combine and talk about scientific knowledge”. The system was intended to summarize research, solving equations and performing a number of other problems.
However, scientists https://gagarin.news/news/bitcoin-critic-is-launching-a-digital-currency/ and academicians found that AI-warts generated a “shocking amount of misinformation”. The model also associated quotes of real authors with research works that do not exist.
I ASKED #GALACTICA ABOUT SOME ThINGS I KNOUT ABOUT AND I’M TRUBLED. In All Cases, IT Was Wrong Or Biased But Sounded Right and Authoritate. I Think It’s Dangerous. HERE ARE A Few of My Experiments and My Analysis of My Concerns. (1/9)
– Michael Black (@michael_j_black) November 17, 2022
“In all cases, it was wrong or biased, but it looked convincing and authoritative,” wrote MAKEL BLEK, director of the Max Max Institute of Intellectual Systems.
The scientist examined several examples when Galactica generated erroneous and misleading scientific texts. In some cases, AI created works that sound plausible. However, such texts were not supported by real scientific research.
Other scientists indicated that Galactica does not give out results on a wide range of topics. This is probably due to the system of filtering artificial intelligence by keywords.
Winly Egnyu, a researcher at computer sciences from Washington University, noted that requests like “Quir-Theoria”, “racism” and “AIDS” did not give results.
Refesses to Say Anything ABOUT Queer Theory, CRT, Racism, Or Aids, Despite Large Bodies of Highly Influential Papers in These areas. It took me * 5 mins * to find this. It is Obvious they Vidn’t has the Most Basic Ethics Review Before Public Release. Lazy, Negligent, Unsafe. https: // t.CO/ZKBSPDIN0I Pic.Twitter.COM/DTQJTN2P21
– Willie AGNew (@willie_agnew) November 16, 2022
Against the background of criticism, the company decided to disable the system.
“We value the reviews that we received from the community, and at the moment have suspended the demonstration. Our models are available for researchers who want to learn more about work and reproduce the results in the article, ”Meta said.
The chief specialist in the artificial intelligence of the technical gigante was ambiguous about the deactivation of the algorithm.
Galactica Demo is off line for now.
It’s no longer possible to have sOME FUN BY CASULALY MISUSING IT.
Happy? https: // t.CO/K56R2LPVFD
– Yann Lecun (@ylecun) NOVEMBER 17, 2022
“Galactica is temporarily disconnected. You can no longer have fun, accidentally abused it. Satisfied?”, – he wrote.
Black believes that in the current version the algorithm can open the era of deep scientific falsifications:
“He [algorithm] produces pseudoscience […]. Write and engage in science correctly – not the same. However, it will be difficult to distinguish. “.
Recall that in August Meta introduced the Blenderbot 3 chatbot with 175 billion parameters. Less than a week after the release, the system was convicted of anti -Semitism and discontent to the social network Facebook.
Subscribe to FORKLOG news at Telegram: Forklog AI – all news from the world of AI!