Artificial intelligence requires careful regulation – 5/6/2023 – Thomas L. Friedman

For the Merriam-Webster dictionary, a Pandora’s box can be “anything that looks ordinary but produces unpredictable harmful results”. I’ve been thinking a lot about Pandora’s boxes, because we Homo sapiens are doing something we’ve never done before: lifting the lids off two giant Pandora’s boxes at the same time, with no idea what might come out.

One is labeled artificial intelligence and exemplified by names like ChatGPT, Bard and AlphaFold, which testify to humanity’s ability, for the first time, to manufacture something that almost godlike approaches general intelligence, far exceeding the intelligence with which we evolve naturally.

Pandora’s other box is labeled climate change, and with it we humans are for the first time conducting ourselves in a godly way from one climate epoch to the next. Until now, this power was largely limited to natural forces involving Earth’s orbit around the sun.

For me, the big question, as we simultaneously lift the lids, is: what kind of regulations and ethics should we put in place to manage what comes out of them screaming?

Let’s be real: we didn’t understand how much social media would be used to undermine the twin pillars of any free society – truth and trust. So if we approach generative AI with the same recklessness – if we again follow Mark Zuckerberg’s bold mantra in the early days of social networking, “move fast and break things” – oh baby, let’s break things faster, with more strength and depth than anyone can imagine.

“There was a failure of the imagination when social media was unleashed, and then a failure to respond responsibly to its unforeseen consequences after it permeated the lives of billions of people,” Dov Seidman, founder and president of HOW, told me. Institute for Society and LRN.

“We wasted a lot of time – and our way– in a utopian thinking that only good things could come from social networks, just by connecting people and giving them a voice. We cannot allow similar failures with artificial intelligence.”

Therefore, there is “an urgent imperative that these artificial intelligence technologies be used only to complement and elevate what makes us human and unique: our creativity, our curiosity and our best selves, the capacity for hope, ethics, empathy, courage and collaboration with others,” added Seidman.

“The adage that great power comes with great responsibility has never been truer. We cannot allow yet another generation of technologists to proclaim their ethical neutrality and tell us ‘hey, we’re just a platform,’ when these AI technologies are enabling exponentially most powerful and profound forms of human empowerment and interaction.”

For these reasons, I asked James Manyika, head of Google’s technology and society team, as well as Google Research, for his thoughts on the promise and challenge of AI.

“We have to be bold and responsible at the same time,” he replied. “The reason we are bold is that, in so many different domains, AI has the potential to help people with everyday tasks and tackle some of humanity’s greatest challenges – like health – and make new discoveries and scientific innovations and productivity gains. that lead to greater economic prosperity.”

It will do this, he added, by “giving people everywhere access to the sum of the world’s knowledge – in their own language, in their preferred mode of communication, via text, speech, images or code”, delivered by smartphone, by television, radio or e-book. Many more people will be able to get better care and better answers to improve their lives.

But we must also be responsible, added Manyika, citing several concerns. First, these tools need to be fully aligned with humanity’s goals. Second, in the wrong hands, they can do enormous damage, whether we’re talking about misinformation, perfectly faked stuff or hacking. The bad guys are always the first adopters.

Finally, “engineering is ahead of science to some extent,” explained Manyika. That is, even the people who build the great language models don’t fully understand how they work and the full extent of their capabilities.

In other words, we still don’t understand how many good or bad things these systems can do. So we need some regulation, but it has to be done carefully.

Why? Well, those most concerned about China overtaking the US in AI want to supercharge AI innovation, not slow it down. If you want to democratize AI, you might want to open source it. But open source can be exploited. What would the Islamic State group do with the code? If you’re worried about AI systems increasing discrimination, privacy violations, and other societal harms, you want regulations now.

If you want to take full advantage of the expected productivity gains from AI, you must focus on creating new opportunities and safety nets for all the researchers, consultants, translators and workers who may be replaced today, and perhaps lawyers and programmers tomorrow. If you’re worried that AI will become super-intelligent and start setting its own goals, regardless of human harm, you should stop it.

This last danger is so real that, on the last day of the 1st, Geoffrey Hinton, one of the pioneers in the creation of AI systems, announced that he is leaving the AI ​​team at Google. He said he thought Google was acting responsibly in launching its AI products, but he wanted to be free to talk about any risks. “It’s hard to see how you can keep bad actors from using it for bad things,” he said.

As a society, we are about to have to decide on some very big tradeoffs as we embrace generative AI.

And government regulation alone will not save us. I have a simple rule: the faster the pace of change and the more seemingly godlike powers we humans develop, all that is old and slow matters more than ever.

The more we scale artificial intelligence, the more the golden rule will need to be reinforced: do unto others as you would have them do unto you. Given the divine powers we are endowing ourselves with, we can now make each other faster, cheaper, and deeper than ever before.

The same goes for the climate Pandora’s box we are opening. As NASA explains on its website, “Over the past 800,000 years, there have been eight cycles of ice ages and warmer periods.” The last ice age ended around 11,700 years ago, giving way to our current climatic age –known as the Holocene–, characterized by seasons that allowed for stable agriculture, the building of human communities and, finally, civilization as we know it. today.

“Most of these climate changes are attributable to very small variations in Earth’s orbit that alter the amount of solar energy our planet receives,” notes NASA. Well, say goodbye to that. There is now an intense discussion among environmentalists about whether we humans have moved out of the Holocene into a new era called the Anthropocene.

This name comes “from anthropo, meaning ‘man’, and ‘cene’, from ‘new’, because humanity has caused mass extinctions of species, polluted the oceans and altered the atmosphere, among other lasting impacts”, as explained by one article in Smithsonian Magazine.

Earth system scientists fear that this man-made epoch, the Anthropocene, will have none of the predictable seasons of the Holocene. Farming could become a nightmare.

But this is where AI can be our savior – accelerating advances in materials science, battery density, fusion energy and safe modular nuclear power that allow humans to manage the impacts of climate change that are now inevitable and avoid those that would be. uncontrollable.

But if AI gives us a way to cushion the worst effects of climate change, we’d better do it right. That means smart regulations to rapidly scale up clean energy with reinforced sustainable values.

Unless we spread an ethic of conservation, we could end up in a world where people feel entitled to drive through the woods now that their Hummer is electric. This can not happen.

In short: these two big Pandora’s boxes are being opened. God save us if we acquire divine powers to part the Red Sea but fail to enforce the Ten Commandments.