‘Godfather of AI’ exits tech world, warning of threat to humanity
Author : Sophia Bell, Group Editor, Connectivity
02 May 2023
(Image: Shutterstock)
A life-long artificial intelligence advocate has announced his resignation from Google as a result of concerns about the dangers of language models such as ChatGPT.
Geoffrey Hinton, 75, is a British-Canadian cognitive psychologist and computer scientist, who is known in the AI community for co-authoring a seminal paper with David Rumelhart and Ronald Williams on the concept of ‘backpropagation’ in 1986. This marked a huge milestone on the journey to widespread AI, popularising backpropagation as a method of training artificial neural networks to mimic human brains in the way they process information.
Today, this so-called ‘deep learning’ underpins almost all machine learning models, including ChatGPT, allowing them to learn from their mistakes.
Hinton’s work on deep learning has earnt him the moniker ‘the Godfather of AI’, as well as the Turing Award in 2018, the world’s top award in computer science, with two of his collaborators.
A life-long advocate for AI, Hinton has now proclaimed his resignation from Google (where he has worked for over a decade) and from the tech world more generally, citing concerns around the threat that the technology he helped to pioneer poses to humanity.
Although AI is by no means a new technology, the rapidity with which it is developing is causing alarm among experts such as Hinton. "The idea that this stuff could actually get smarter than people – a few people believed that," Hinton told The New York Times.
"But most people thought it was way off. And I thought it was way off. I thought it was 30 to 50 years or even longer away. Obviously, I no longer think that."
Hinton’s warning comes against a background of recent rise in language models such as the headline-grabbing ChatGPT and its recently announced successor, GPT-4. The latter not only already outperforms ChatGPT in writing text, but also has an advanced capacity to recognise and explain images.
It also has Hinton to thank, in part, for its development. In 2012, Hinton collaborated with two other researchers (including Ilya Sutskever, who later founded OpenAI) to build a computer vision system that could recognise hundreds of objects in pictures, an ability that GPT-4 has taken to new heights.
With AI’s capabilities accelerating, more and more experts are becoming wary of its implications on civilisation. At the end of March, more than 1,000 industry leaders, including Elon Musk and Professor Stuart Russell, signed an open letter which called for a six-month pause in the development of such systems.
"AI systems with human-competitive intelligence can pose profound risks to society and humanity," the letter warned.
"Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable."
The letter urged the need to put regulations in place to try to ensure a safer approach to the technology’s development and to mitigate the profound risks it poses to society. "AI labs and independent experts should use this pause to jointly develop and implement a set of shared safety protocols for advanced AI design and development that are rigorously audited and overseen by independent outside experts," it stated.
"These protocols should ensure that systems adhering to them are safe beyond a reasonable doubt."
The feeling from these experts is that tech giants and governments alike are not doing enough to keep the technology in check. For example, the UK Government’s “pro-innovation approach” outlined in its National AI Strategy, has recently been criticised for its “light touch”, and for not proposing any new legislation that will directly and practically tackle the issue.
The difficulty of controlling AI and the dangers of disinformation are among the most cited reasons for greater regulation. Its potential to make many roles redundant is also causing disquiet for many. Although the introduction of robotics over the past couple of decades has also sparked concerns about job security, on the whole, industry has welcomed automation for its ability to allow human workers to cast off the dull, dirty and dangerous, in favour of more stimulating and rewarding tasks.
A similar argument could be leveraged in favour of artificial intelligence; however, Dr Hinton fears that its ability may not be limited simply to a complementary role in the workplace: “It takes away the drudge work,” he said. “It might take away more than that.”