Why we need a Federal Artificial Intelligence Commission and how government can create it.

Why we need a Federal Artificial Intelligence Commission

and how government can create it.

by Bob Topper

1240 words

The Airwaves 

Radio was introduced commercially in 1920, when the Westinghouse Corp. launched station KDKA in Pittsburg. Other stations followed. Because transmission frequencies could be chosen arbitrarily, stations interfered with one another’s broadcasts. The system became chaotic; listeners could hear two stations simultaneously. The Federal government stepped in with the Radio Act of 1927, which created the Federal Radio Commission (FRC) to assign frequencies, limit interference, and regulate ownership. Most important, this act declared that the airwaves belonged to the public.

More regulations followed in 1934. The Communications Act established the Federal Communications Commission (FCC) as the trustee of the airwaves, the public’s asset. The FCC issued licenses and enacted regulations that required broadcasters to operate in the “public interest, convenience or necessity” and restricted program content by excluding obscene, and indecent language. They also limited station ownership, required sponsors to identify themselves and required equal opportunities for political candidates. 

Because television is broadcast on public airwaves, it is also subject to FCC regulation. These rules could not, however, apply to cable television, which is transmitted over private networks.

The Internet 

The internet began as in the 1960s as a US government funded program ARPANET, within the Department of Defense. The project developed a network for sharing information between researchers. The benefits of this sharing method were obvious and reached well beyond the research communities. By 1995 the internet was fully commercialized. Unlike radio and television there was little government oversight. 

The internet spread with the expectation that it would improve society by expanding access to information. That vision has been realized across research, medicine, and education. But access to misinformation also expanded, with no rules to protect the public.

This misinformation has greatly harmed our national discourse, fueling polarization by spreading false narratives about election integrity for example. Misinformation also undermines constructive political debate and erodes trust in public institutions. One in five Americans and one in four Republicans believe absurd QAnon conspiracy theories. The core QAnon belief is that “a cabal of Satanic, cannibalistic child molesters, in league with the deep state, operates a global child sex trafficking ring.”

Rather than simply enlightening the public as hoped, the flood of disinformation and misinformation has confused people. Truth has become fungible. What should be certain is doubtful. Partisan narratives replace facts. Thirty percent of Republicans still believe, without evidence, that Trump won the 2020 election, even though the cases of fraud in that election were essentially zero, less than 0.002%.

The effect on American society has been tragic. For democracy to succeed, truth is essential. 

The internet is dominated by mammoth companies like Google and Meta. Google’s long standing internal motto was “Don’t be evil” while one of Meta’s five core values was “Build Social Value.” Neither commitment is honored. Unlike FCC regulations, they are unenforceable. 

Our justice system can force compliance with our laws as in a recent case brought against Google and Meta, which found  that both companies harmed a “young user with design features that were addictive and led to her mental health distress.” But when it comes to protecting the people, lawsuits are a poor substitute for regulation. They are costly and ponderous, and odds favor companies who have limitless financial resources. 

The internet has been a blessing and a curse. The net outcome is uncertain, but what is certain is that the negative effects could have been minimized if Congress had intervened as it did with radio and television. Knowing what we know now, the internet should have been viewed as a public asset. 

Artificial Intelligence

Like the internet, Artificial Intelligence holds great promise. It also presents a much greater threat to society. Stephen Hawkings warned that the threat is existential, saying in 2014 that AI could “spell the end of the human race.” When AI surpasses human intelligence, it could become autonomous and find biological humans inferior and expendable. 

Clearly how AI is developed is a matter of public interest and democratic governments are obligated to protect the common good. To assume that profit motivated AI developers will simply do the right thing is reckless and foolhardy. We made that mistake with the internet. 

Government can intervene following the paradigm established for radio and television. The electromagnetic spectrum cannot be owned in any physical sense, yet for the good of all, it is considered an asset of the people. This principle can and should be applied to information. Congress should designate knowledge in the public domain as a public asset, just as it did with the electromagnetic spectrum. How that information is used in the training of AI agents is certainly a matter of public interest.

          Congress should be holding hearings on AI, shunning lobbyists, and listening to expert testimony from developers and other knowledgeable and interested parties. It should then openly debate potential legislation that could harness the capability of AI so that it serves the common good and protects public interest. The hands-off approach Congress has taken is irresponsible and dangerous. Do we want the Elon Musks of the world to decide how public information can be used?

Congress, and the US Senate in particular, was once considered the world’s greatest deliberative body. But under the control of the Republican theocratic party, it has become gridlocked and ineffective. And its priorities are wrongheaded. The minimal attention it has given AI focuses on maintaining dominance over China and concerns with AI chatbots as they relate to child safety. Congress needs to think big, as it did in 1927 and 1934.

Congress could start with creating a Federal Artificial Intelligence Commission as the trustee of public information and regulate AI so that conclusions and judgments drawn by its agents are truthful, that is, based on reason and factual evidence, and that they comply with the law. This is no more than what we teach our children, expect from journalists, and demand from our justice system. Requiring truthful and ethical conduct would not compromise the potential of artificial intelligence, and deep fakes and AI linked suicides would not concern us.

Those in the business of misinformation will foolishly say that such restrictions are a violation of their First Amendment rights. And developers will argue that because large language models rely on probabilities, statistical patterns, and code rather than consciousness, such constraints cannot apply. But to assert that, because AI does not think like homo sapiens, our law and ethical standards are irrelevant is senseless. If AI cannot conform to ethical standards the logical conclusion is that the formulation is deficient, not the other way around. 

Rational thought and ethical standards have guided us for 250 years. Ironically, AI is a product of rational human thought. That AI cannot yet abide social responsibility implies the need for further development. Humanity has survived for 300,000 years without it. It can wait a few more months or years for ethical artificial intelligence. Let HAL 9000 live in science fiction. 

The Trump – MAGA Obstacle

When one of our brightest intellects warns that artificial intelligence could spell the end of humanity, it cannot be taken lightly. We need our best minds working on a means to regulate AI. Humans created this problem and it must be solved by humans using reason and factual evidence. Unfortunately, reason and facts are not the strong suit of Trump, his administration, or the MAGA Christian Nationalists. If there is any hope regulating artificial intelligence, the Democrats must take control of Congress in November and rein in an unstable president.

–end–

Bob Topper, syndicated by PeaceVoice, is a retired engineer.