Should we really fear AI?

The take on Geoffrey Hinton’s resignation from Google after being employed for more than a decade has been pretty much straightforward, with the elderly British scientist plainly clear about the whys as gleaned from his succeeding interviews and Twitter account.

“Actually, I left so that I could talk about the dangers of AI without considering how this impacts Google. Google has acted very responsibly,” Hinton quickly replied on Twitter last May 1 to New York Times technology reporter Cade Metz’s implied statement that his resignation was in order to criticize his former employer.

Hinton’s fear of artificial intelligence (AI), particularly the generative type that burrows on deep learning neural networks, is that it can get unhinged and fly into territories that would make it dangerous to mankind.

During an interview with Will Douglas Heaven, senior editor for AI at the MIT Technology Review, Hinton stated his reasons for resigning from Google. “One was that I’m 75, and I’m not as good at doing technical work as I used to be. My memory is not as good and when I program, I forget to do things. So it was time to retire.

“A second was, very recently, I’ve changed my mind about the relationship between the brain and the kind of digital intelligence we’re developing. I used to think that the computer models we were developing weren’t as good as the brain. The aim was to see if you could understand more about the brain by seeing what it takes to improve the computer models.

“Over the last few months, I’ve changed my mind completely, and I think probably the computer models are working in a completely different way than the brain. They’re using back propagation and I think the brain’s probably not. And a couple things have led me to the conclusion and one of them is the performance of GPT-4.”

Better than humans

Back propagation, an algorithm that was developed by Hinton to train computers to develop its “reasoning” ability, has bred what is now popularly known as deep learning, and which has earned Hinton the title of being the godfather of AI.

With recent breakthroughs in computer deep learning, the world has been seeing quantum leaps in knowledge accumulation by computers. ChatGPT, and many other chatbots that have been introduced in the public arena towards the end of 2022, are developing at a rate that Hinton fears will soon surpass human intelligence, but more importantly, may be used in ways that would harm people.

Hinton warns that chatbots “will have learned from us by reading all the novels that ever were and everything Machiavelli ever wrote [about] how to manipulate people. And if they’re much smarter than us, they’ll be very good at manipulating us.

“You won’t realize what’s going on. You’ll be like a two-year-old who’s being asked, ‘Do you want the peas or the cauliflower,’ and doesn’t realize you don’t have to have either. And you’ll be that easy to manipulate.

“They can’t directly pull levels, but they can certainly get us to pull levers. It turns out if you can manipulate people, you can invade a building in Washington without ever going there yourself.”

Unfortunately, while Hinton has set off an alarm to warn us, he can offer no easy way out. In fact, he admits that there might not be any solution. “But I think it’s very important that people get together and think hard about it, and see if there’s a solution.”

He goes on to say that, “So, the good news is we figured out how to build beings that are immortal. When a piece of hardware dies, they don’t die. If you’ve got the weights stored in some medium and you can find another piece of hardware that can run the same instructions, then you can bring it to life again.

“So, we’ve got immortality, but it’s not for us.”

Pragmatic view

If that’s the view of Hinton as AI’s godfather, AI’s acknowledged father poses a different perspective. Jürgen Schmidhuber, whose work on neural networks became the basis for such technologies as Google Translate and Apple’s Siri, dishes out a more pragmatic view.

In an interview on The Guardian, Schmidhuber describes the progress on AI as something that cannot be stopped, given the extent by which countries and companies are racing to get ahead of one another.

Furthermore, he believes that development of AI should not be halted: “But then I think you also shouldn’t stop it. Because in 95 percent of all cases, AI research is really about our old motto, which is make human lives longer and healthier and easier.”

Schmidhuber posits that the way to deal with the bad guys who use AI for misguided purposes is to counter by developing good tools that would put a stop to the bad guys. Unfortunately, this view is not shared by many of his colleagues.

Pause for guardrails

An open letter issued on March 22, which has now garnered over 27,000 signatures from vested stakeholders, calls for a pause in all giant AI experiments with human-competitive intelligence for six months because of “profound risks to society and humanity, as shown as extensive research and acknowledged by top AI labs.”

Seven policy recommendations were called for, including the mandating of third-party auditing and certification down to the development of standards for identifying and managing AI-generated content and recommendations.

Whether someone of influence can really act on the open letter and call for a pause in this AI arms race is highly debatable. In all probability, it is no longer within anyone’s control.

Facebook and Twitter

We are actively using two social networking websites to reach out more often and even interact with and engage our readers, friends and colleagues in the various areas of interest that I tackle in my column. Please like us on www.facebook.com/ReyGamboa and follow us on www.twitter.com/ReyGamboa.

Should you wish to share any insights, write me at Link Edge, 25th Floor, 139 Corporate Center, Valero Street, Salcedo Village, 1227 Makati City. Or e-mail me at [email protected]. For a compilation of previous articles, visit www.BizlinksPhilippines.net.