The tech community is in a frenzy over an open letter signed by some of the industry’s biggest names warning of the civilizational risks posed by the development of artificial intelligence.
The letter proposes a six-month moratorium on “the training of AI systems,” and it describes some alarming scenarios that could be in store for us if their warnings are ignored.
Should we let machines flood our information channels with propaganda and untruth? Should we automate away all the jobs, including the fulfilling ones? Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization? Such decisions must not be delegated to unelected tech leaders.
Given the quasi-governmental powers arrogated and jealously stewarded by the tech industry over the years, the long-overdue deference that the last sentence offers to elected lawmakers is welcome. That impulse becomes somewhat less laudable when the reader realizes that it serves only to grease the skids for a variety of fund-raising solicitations.
The letter concludes with some recommendations, including “new and capable regulatory authorities dedicated to AI” (which, because they would be new, are especially at risk of regulatory capture). The authors demand “robust public funding for technical AI safety research” and seek “well-resourced institutions for coping with the dramatic economic and political disruptions (especially to democracy) that AI will cause.”
In sum, AI demands a bigger government in Washington, more substantial taxpayer-funded outlays for “research,” and more industrial protectionism lest we risk losing our “democracy.” If an AI had written this appeal to the fixations that haunt the imaginations of the profligates who populate Washington, it couldn’t have done much better.
The letter’s apocalyptic techno-pessimism has not been well received even by techno-pessimists. As Axios reported, technology professor and AI critic Emily Bender was unimpressed by the “hype” that she saw “dripping” from the letter. “I’m glad that the letter authors [and] signatories are asking ‘Should we let machines flood our information channels with propaganda and untruth?’” She continued: “But the questions after that are just unhinged #AIhype, helping those building this stuff sell it.” Other critics of the letter note that it elides the distinctions between modern AI and artificial general intelligence, which would approach a human’s capacity for cognition and is a technology that’s decades away — if it is feasible at all.
Implementing basic risk-mitigation strategies associated with this new technology does not require a six-month moratorium on AI research, even if that was a desirable proposition. (It is not, in part, because to do so would be to unilaterally cede the field to foreign commercial competitors and foreign militaries.) Semiautonomous, artificially intelligent weapons platforms and cyberwar capabilities may still be years away, but their development would have a deterrent effect like any other capability. Sacrificing those capabilities sacrifices deterrence.
Moreover, as technological revolutions of the past have repeatedly proven, the threat that AI poses to our present way of life is eclipsed by the social and economic benefits it promises. The more routinized a task is, the more likely AI and automation will replace it within the next two decades. That will be disaggregating and disorienting, and there will be political dysfunction as a result. But trying, as the political classes of the past have, to protect the vulnerable from the ravages of innovation does not put a stop to technological progress.
To hear advocates of protectionism tell it, protectionism has routinely failed to achieve its primary goal: protecting workers in industries threatened by technological advances. Nor have their worst-case predictions come to pass. Because of technological innovation, overall productivity and economic growth have increased, and living standards even for America’s poorest have improved markedly.
Even AI optimists aren’t dismissive of the notion that something approaching sentience in machines could be a dangerous development. But the dangers are evitable. And all attitudes toward it assume that this technological feat is achievable, which may be a flawed assumption.
“Every time we make progress,” University of Michigan engineering and computer-science professor John Laird told Built In magazine reporters, “we also get a new appreciation for how hard it is.” Northwestern University’s Diego Klabjan isn’t sure we’ll ever develop something that rivals the human brain. “The current state-of-the-art [technology] is just straightforward connections following very easy patterns,” he observed. “So going from a few million neurons to billions of neurons” to match the computational power of the human mind “with current hardware and software technologies — I don’t see that happening.”
All this humility is absent from the open letter setting the tech world on fire. The certainty it expresses in its predictions is matched only by its conviction that the AI menace won’t be contained until and unless you pony up. The letter’s authors wanted you to reach for your shotgun. You should be reaching for your wallet instead.