Elon Musk says his synthetic intelligence firm xAI will retrain its AI mannequin, Grok, on a brand new data base freed from “rubbish” and “uncorrected information” — by first utilizing it to rewrite historical past.
In an X submit on Saturday, Musk stated the upcoming Grok 3.5 mannequin could have “superior reasoning” and wished it for use “to rewrite your entire corpus of human data, including lacking data and deleting errors.”
He stated the mannequin would then retrain on the brand new data set, claiming there was “far an excessive amount of rubbish in any basis mannequin skilled on uncorrected information.”
Musk’s newest battle in opposition to “woke”
Musk has lengthy claimed that rival AI fashions, similar to ChatGPT from OpenAI, a agency he co-founded, are biased and omit data that isn’t politically right.
For years, Musk has regarded to form merchandise to be free from what he considers to be damaging political correctness and has aimed to make Grok what he calls “anti-woke.”
He additionally relaxed Twitter’s content material and misinformation moderation when he took over in 2022, which noticed the platform flooded with unchecked conspiracy theories, extremist content material and pretend information, a few of which was unfold by Musk himself.
Musk aimed to battle the tide of misinformation by implementing a “Group Notes” function, permitting X customers to debunk or add context to posts that present prominently underneath offending posts.
Criticism levelled at Grok’s retraining
Musk’s submit attracted condemnation from his critics, together with from Gary Marcus, an AI startup founder and New York College professor emeritus of neural science who in contrast the billionaire’s plan to a dystopia.
“Straight out of 1984,” Marcus wrote on X. ”You couldn’t get Grok to align with your individual private beliefs so you will rewrite historical past to make it conform to your views.”
Bernardino Sassoli de’ Bianchi, a College of Milan professor of logic and science philosophy, wrote on LinkedIn that he was “at a lack of phrases to touch upon how harmful” Musk’s plan is.
Associated: Want ChatGPT’s assist? It could possibly be making you dumber
“When highly effective billionaires deal with historical past as malleable just because outcomes don’t align with their beliefs, we’re not coping with innovation — we’re dealing with narrative management,” he added. “Rewriting coaching information to match ideology is fallacious on each conceivable stage.”
Musk’s name for “info” brings conspiracy theories, falsehoods
As a part of his effort to overtake Grok, Musk known as on X customers to share “divisive info” to coach the bot, specifying they need to be “politically incorrect, however nonetheless factually true.”
The replies noticed quite a lot of conspiracy theories and debunked extremist claims, together with Holocaust distortion, debunked vaccine misinformation, racist pseudoscientific claims relating to intelligence and local weather change denial.
AI Eye: AI is nice for employment says PWC — Ignore the AI doomers