Is the US government ready for the rise of artificial intelligence?
We’re at a Frankenstein moment. An artificial intelligence boom is taking over Silicon Valley, with hi-tech firms racing to develop everything from self-driving cars to chatbots capable of writing poetry.
Yet AI could also spread conspiracy theories and lies even more quickly than the internet already does – fueling political polarization, hate, violence and mental illness in young people. It could undermine national security with deepfakes.
In recent weeks, members of Congress have sounded the alarm over the dangers of AI but no bill has been proposed to protect individuals or stop the development of AI’s most threatening aspects.
Most lawmakers don’t even know what AI is, according to Representative Jay Obernolte, the only member of Congress with a master’s degree in artificial intelligence.
What to do?
Many tech executives claim they can simultaneously look out for their company’s interests and for society’s. Rubbish. Why should we assume that their profit motives align perfectly with the public’s needs?
Sam Altman – the CEO of OpenAI, the company responsible for some of the most mind-blowing recent advances in AI – believes no company, including his, should be trusted to solve these problems. The boundaries of AI should be decided, he says, not by “Microsoft or OpenAI, but society, governments, something like that”.
But does anyone trust the government to do this? If not, how can “society” manage it? Where can we look for a model of how to protect ourselves from the downsides of an emerging technology with such extraordinary upsides, without stifling it?
One place to look is Herbert Hoover. Seriously. Not when Hoover was president and notoriously failed to do anything about the Great Depression, but when he was US secretary of commerce between 1921 to 1929.
One of Hoover’s great achievements a century ago, largely unrecognized and unremembered today, was managing the development of a new and crucial technology in the public interest.
That new technology was electricity. Thomas Edison and other entrepreneurs and the corporations they spawned were busily promoting all manner of electric gadgets.
Those gadgets had the potential to make life easier for millions of people. But they could also pose grave dangers. They could destroy buildings, and injure or kill people.
Hoover set out to ensure that the infrastructure for electricity – wires, plugs, connectors, fuses, voltage and all else – was safe and reliable. And that it conformed to uniform standards so products were compatible with one another.
He created these standards for safety, reliability and compatibility by convening groups of engineers, scientists, academics, experts and sometimes even journalists and philosophers – and asking them to balance public and private interests. He then worked with the producers of electric gadgets to implement those standards.
Importantly, the standards were non-proprietary. No one could own them. No one could charge for their use. They were, to use the parlance of today, “open source”.
Much of today’s internet is based on open-source standards. We take them for granted. Computers could not communicate without shared models, such as HTTP, FTP and TCP/IP.
Although digital standards haven’t protected the public from disinformation and hate speech, they have encouraged the creation of services such as Wikipedia, which are neither privately owned nor driven by profits.
In fact, you could view our entire system of intellectual property – copyrights, patents and trade names – as premised on eventual open-source usage. After a certain length of time, all creations lose their intellectual property protections and move into the public domain where anyone is free to use them. (Not incidentally, when he was secretary of commerce, Hoover advanced and streamlined the intellectual property system.)
So what would Hoover have done about AI?
He wouldn’t wait for the producers of AI to set its limits. Nor would he trust civil servants to do it. Instead, he’d convene large and wide-ranging panels to identify AI’s potential problems and dangers, come up with ideas for containing them, and float the ideas with the public.
If the proposed standards stood the test, he’d make them voluntary for the industry – with the understanding that the standards could be modified if they proved impracticable or unnecessarily hobbled innovation. But once in place, if corporations chose not to adapt the standards, their AI products would lose intellectual property protections or be prohibited.
Hoover would also create incentives for the creation of open-source AI products that would be free to the public.
In other words, Hoover wouldn’t rely solely on business or on government, but on societyto gauge the common good.
AI has the potential for huge societal benefits, but it could also become a monster. To guide the way, we need the leadership and understanding of someone like Herbert Hoover when he was secretary of commerce.
Robert Reich, a former US secretary of labor, is professor of public policy at the University of California, Berkeley, and the author of Saving Capitalism: For the Many, Not the Few and The Common Good. His new book, The System: Who Rigged It, How We Fix It, is out now. He is a Guardian US columnist. His newsletter is at robertreich.substack.com