
There is something quite bizarre about the current debate on artificial intelligence. Never before have the heads of major corporations come out with warnings that their own products could threaten the future of the human race.
Such statements cannot be taken at face value. If Dario Amodei, Sam Altman, and Elon Musk believe those catastrophic possibilities to be literally true, they would be talking about dismantling AI, not simply slowing down its progression.
I’m not suggesting that the risks of AI are entirely fabricated, but the industry leaders seem to be using worst-case scenarios to try to throw us into a panic. The question is why.
One possibility is they are seeking to promote the idea that the dangers are so great that only they are qualified to handle them. This would be a bold new form of regulatory capture. The implied message is that only the geniuses who built AI are in a position to try to control it.
The AI barons may thus be setting the stage for a system of industry self-regulation that will not seriously inhibit their actions but give the impression they are addressing the risks. This would be similar to the mostly toothless measures adopted by social media companies in response to concerns about the harms to young people.
It is true that the AI leaders have made reference to government involvement, but it is likely they are doing so on the assumption that any such intervention will be less than robust or will not materialize at all. Donald Trump reinforced that supposition by immediately coming out in opposition to any new controls on AI, saying all that is needed is a “strong and smart (high IQ!) president.”
Another possibility is that the big AI companies are serious about wanting a slowdown, but they would seek to shape it so that it solidifies their dominant position by making it more difficult for competitors to emerge. The adoption of new guardrails on the technology would make it much more expensive to develop new AI products and thus price out upstart companies.
The AI industry may also be promoting cataclysmic thinking as a way of inoculating themselves against future legal liability. They could be assuming that by issuing their warnings now, they will be blamed less later when serious, but not quite extinction-level, consequences begin to materialize. We warned you, they may anticipate saying, and you failed to take appropriate action.
By offering up the worst possible scenarios, the industry may be seeking to numb people to the risks. Looking at what has happened with the climate crisis, they may be thinking that all the talk of catastrophe and human extinction will cause many to adopt a posture of hopelessness or indifference.
These theories of AI industry motivation may seem cynical, but it is difficult to believe that companies claiming to be capable of producing superintelligence are truly helpless to control their own creations.
Readers in DC may be interested to know that Michael Marx, author of Reining in the Bulls, the great new book on corporate campaigns I wrote about last month, will be appearing on October 1 at 6 PM at the Busboys and Poets at 450 K Street NW. The event will include a discussion on corporate accountability with Kathy Mulvey, Climate Accountability Campaign Director at the Union of Concerned Scientists. You can register here.








You must be logged in to post a comment.