The same AI researcher who said AGI (Artificial General Intelligence) will “solve all the problems that we have today” before warning that they will also present “the potential to create infinitely stable dictatorships.” has posted a Tweet that has set off a considerable debate in the scientific community.
Ilya Sutskever On February 9, Tweeted:
it may be that today’s large neural networks are slightly conscious— Ilya Sutskever (@ilyasut) February 9, 2022
Noor Al-Sibai who writes for futurism.com says ‘Needless to say, that’s an unusual point of view. The widely accepted idea among AI researchers is that the tech has made great strides over the past decade, but still falls far short of human intelligence, nevermind being anywhere close to experiencing the world consciously.’
Many scientists treat Sutskever’s claim with a great degree of skepticism because Ilya Sutskever, the head scientist at Elon Musk, cofounded the research group OpenAI. They are comfortable dismissing Sutskever’s claims as a sensationalist “sales pitch” for OpenAI’s work. Some are incredibly annoyed that such fantastical claims about the capabilities of AI at present are harming machine learning’s reputation.
Prominent Meta backed expert LeCunn disagrees with claim
First to wade in and criticise the claim was Chief AI Scientist at Meta, Yann LeCunn who tweeted in response:
Nope.— Yann LeCun (@ylecun) February 12, 2022
Not even for true for small values of “slightly conscious” and large values of “large neural nets”.
I think you would need a particular kind of macro-architecture that none of the current networks possess.
There has been a longstanding spat between Musk and Zuckerberg, so it maybe shouldn’t be a surprise Meta’s AI chief scientist was first to wade in a claim by a prominent Musk-backed expert.
Then Turing Award winner Judea Pearl agreed with Yann LeCunn, tweeting, “Before a system can lay claims to consciousness, it must exhibit “deep understanding” of some domain, which large NN’s have yet to exhibit by answering questions at all three levels of the reasoning hierarchy.”
MIT expert gives his opinion
At first, it looked like the backlash in the scientific community had isolated Sutskever in his viewpoint, but now, MIT computer scientist Tamay Besiroglu has come to Sutskever’s defense.
Besiroglu sent a series of Tweets in Sutskever’s defence:
“Seeing so many prominent [machine learning] folks ridiculing this idea is disappointing,”
“It makes me less hopeful in the field’s ability to seriously take on some of the profound, weird and important questions that they’ll undoubtedly be faced with over the next few decades.”
“I don’t actually think we can draw a clear line between models that are ‘not conscious’ vs. ‘maybe slightly conscious,’” he said in a follow-up message. “I’m also not sure any of these models are conscious.”
“That said, I do think the question could be a meaningful one that shouldn’t just be neglected,”
Do AI experts agree what consciousness is?
From a thoughtful analysis by Noor Al-Sibai, the real issue of the debate between AI experts is not whether AI is conscious or not as we understand consciousness to be, but that the ‘boundaries of what constitutes consciousness‘ needs to be redefined.
When scientists can agree on the definition of consciousness, only then will we be able to say with absolute certainty whether AI has gained consciousness or not.
Defining what consciousness is…
Jacy Reese Anthis, social scientist, co-founder of the Sentience Institute was happy to define what consciousness is at length in a Twitter thread. He acknowledged it is a ‘deeply debated topic among philosophers and scientists’.
And this very much indicates specific measures are needed to define consciousness because terms/words can be too vague, and he writes that ‘consciousness does not exist objectively.’ In summary, people will never be able to agree on what consciousness is without measures because personal feelings or opinions influence a person’s interpretation of the word consciousness.