Commentary: Artificial Intelligence, Email, Government Oversight, and ... the Printing Press

The significance of the printing press as a societal achievement was recently recognized in a highly complimentary way.

At a Senate Judiciary Subcommittee hearing in Washington, DC on May 16, 2023, Sam Altman, CEO of OpenAI, a San Francisco artificial intelligence (AI) start-up, warned: If AI technology is not regulated by the Federal Government, it could have a negative impact on privacy, quality of life, safety, and legal use of technology. Altman emphasized that for such reasons, his company and others invested in the development of AI technology, such as Google and Microsoft, must support the need for government oversight. If no such oversight is provided, AI technology may not have the disruptive but positive impact that the printing press had, and may have the disruptive, negative, impact of the likes of the atomic bomb.

I hasten to add the following.

I am not advocating that government control and intervention is always a good thing. I am advocating that government control and intervention is sometimes needed and appropriate to protect the rights of The People. Indeed, the printing press was probably the most significant technology protecting and advancing the First Amendment right of freedom of speech, i.e., freedom of the press. Without government intervention, the protection of this right may not have existed, and society recognizes this. For example, at the turn of the 21st century, a think-tank group known as The Edge conducted a study of people in many professions (doctors, attorneys, professors, artists, writers, philosophers, and others). They were asked the question: Of all inventions of the past 2000 years, what had the greatest positive impact on the advancement of society? By a factor of 3 to 1, Gutenberg’s movable type and the printing press was ranked No. 1.

Now, I’ll propose an area of technology that, initially, was controlled by the government, but then control was relinquished to private businesses to do what it cared to with the technology. The technology is email.

Email was first developed within the labs of academia under the control of the Federal Government and financed by a project called The Arpanet (the Arpanet later became the Internet). During the 1980s and 1990s, use of email became common in business, government, universities, industries, and among the general public. Prior to that, email was for the most part not even a vision as a utility or a term with meaning to business, industry, and the public.

Consider the following. On November 1, 1981, the cost of a First Class postage stamp was 20 cents. On January 10, 1999, the cost of a First Class postage stamp was 33 cents. Suppose that before email became a free public utility, we woke up one morning and learned that the cost of First Class mail was being reduced by 90 percent, so long as mail was sent via email. I propose that the government would have been praised and applauded for this. However, once email became a free utility, there would be no turning back with a cost imposed that would be acceptable to the public.

But consider that if there were always a cost to sending email, even just a few pennies, the most negative aspects of email would have not occurred. For example, there would be no spam, there would be no ransomware, there would be minimal random unwanted advertising…because the senders of such email would have to pay per emailed unit, and such abusers likely send out billions of emails per month. The cost to them would be enormous. However, the public having never conceived of free mail would not object when their mailing cost went down, for example, by 90 percent. And to the government’s advantage, the US Postal Service deficit would no longer exist and, possibly, the national deficit would have been reduced if the US Government retained control of email and received the revenue from it, even at a few cents per mailed unit.

This same scenario, or worse, could occur with AI if this technology goes unregulated. The negative impact of email, previously noted, would pale in comparison to unregulated AI.