‘Godfather’ of AI is now having second thoughts – The B.C. Catholic

Until a few weeks ago British-born Canadian university professor Geoffrey Hinton was little known outside academic circles. His profile became somewhat more prominent in 2019 when he was a co-winner of the A. M. Turing Award, more commonly known as the Nobel Prize for computing.

However, it is events of the past month or so that have made Hinton a bit of a household name, after he stepped down from an influential role at Google.

Hintons life work, particularly that in computing at the University of Toronto, has been deemed groundbreaking and revolutionary in the field of artificial intelligence, AI. Anyone reading this column will surely have encountered numerous pieces on AI in recent months, be it on TV, through radio, or in print, physical and digital. AI applications such as large language model ChatGPT have completely altered the digital landscape in ways unimaginable even a year ago.

While at the U of T, Hinton and graduate students made major advances in deep neural networks, speech recognition, the classification of objects, and deep learning. Some of this work morphed into a technology startup which captured the attention of Google, leading to the acquisition of the business for around $44 million a decade ago.

Eventually, Hinton became a Google vice-president, in charge of running the California companys Toronto AI lab. Leaving that position recently, at the age of 75, led to speculation, particularly in a New York Times interview, that he did so in order to criticize or attack his former employer.

Not so, said Hinton in a tweet. Besides his age being a factor, he suggested he wanted to be free to speak about the dangers of AI, irrespective of Googles involvement in the burgeoning field. Indeed, Hinton noted in his tweet that in his view Google had acted very responsibly.

Underscoring his view of Googles public AI work may be the companys slow response to the adoption of Microsoft-backed ChatGPT in its various incarnations. Googles initial public AI product, Bard, appeared months after ChatGPT began its meteoric adoption in early December. It did not gain much traction at the outset.

In recent weeks weve seen news stories of large employers such as IBM serving notice that about 7,000 positions would be replaced by AI bots such as specialized versions of ChatGPT. Weve also seen stories about individuals turning over significant aspects of their day-to-day life to such bots. One person gained particular attention for giving all his financial, email, and other records to a specialized AI bot with a view to having it find $10,000 in savings and refunds through automated actions.

Perhaps it is these sorts of things that are giving Hinton pause as he looks back at his lifes work. In the NYT interview, he uses expressions such as, It is hard to see how you can prevent the bad actors from using it for bad things, and Most people will not be able to know what is true anymore -- the latter in reaction to AI-created photos, videos, and audio depicting objects or events that didnt occur.

Right now, they are not more intelligent than us, as far as I can tell. But they soon may be, said Hinton, speaking to the BBC about AI machines. He went on to add, Ive come to the conclusion that the kind of intelligence we are developing (via AI) is very different from the intelligence we have.

Hinton went on to note how biological systems (i.e. people) are different from digital systems. The latter, he notes, has many copies of the same set of weights and the same model of the world, and while these copies can learn separately, they can share new knowledge instantly.

In a somewhat enigmatic tweet on March 14 Hinton wrote: Caterpillars extract nutrients which are then converted into butterflies. People have extracted billions of nuggets of understanding and GPT-4 is humanitys butterfly.

Hinton spent the first week of May correcting various lines from interviews he gave to prominent news outlets. He took particular issue with a CBC online headline: Canadas AI pioneer Geoffrey Hinton says AI could wipe out humans. In the meantime, theres money to be made. In a tweet he said: The second sentence was said by a journalist, not me, but you wouldnt know that.

Whether the race to a God-like form of artificial intelligence fully materializes, or not, AI is already being placed alongside climate change and nuclear war as a trio of existential threats to human life. Climate change is being broadly tackled by most nations, and nuclear weapons use has been effectively stifled by the notion of mutually-assured destruction. Perhaps artificial general intelligence needs a similar global focus for regulation and management.

Follow me on Facebook (facebook.com/PeterVogelCA), or on Twitter (@PeterVogel)

[emailprotected]

Go here to read the rest:

'Godfather' of AI is now having second thoughts - The B.C. Catholic