Artificial intelligence and moral panics

It has been interesting to observe the moral panic that has emerged around artificial intelligence; much of it accelerated by the appearance of ChatGPT.  Whilst there are different flavours, such as the "AI will destroy our jobs" canard, the most extreme concerns seem to relate to the "AI alignment" problem: the need for AI to align with human values.  

There is, of course, a great deal of work to be done to infuse AI with reason and logic, however some have already concluded that there is no hope.  One  extreme (and somewhat amusing) example of this being the morbid conclusion of the so-called Machine Intelligent Research Institute that there is no point even trying to solve this problem and such efforts should instead be redirected to helping humanity die with dignity in the face of the widespread emergence of super-intelligent artificial general intelligence:

This is, of course, a relatively extreme example: most contemporary debates about "AI safety" are focused less on the existential threat of AI to humanity but rather on how AI can be prevented from engaging in "hate speech", being offensive, or generating misinformation.  

As part of the panic, there are also some people promoting a modern retelling of Luddite fallacies about technology replacing all jobs.  Whereas, "classic Luddism" tended to be concerned with the replacement of blue collar work, it is interesting to see that, with AI, the opposite is likely to be what happens: AI is targeting "creative work" through the automated generation of art and words; and "white collar" work is more likely to be disrupted by AI than blue collar work which is typically much harder to replace.  This panic about job destruction is, of course, not new and goes back to every emerging technology whether it was outsourcing in the 2000s or robots in the 2010s.  The reality is that technology indeed destroys some jobs but it creates many more.  Almost every office job today exists because of some technology innovation that happened in the last 100 years.  

For example, the below graph from US Department of Commerce data clearly shows how technology has led to creation of more jobs than it destroyed:

It is interesting to view the moral panics about artificial intelligence in the context of history's broad arc.  We can find, for example, similar impulses in mythology: the story of Prometheus providing an early example of man's concern with the emergence of new technology.  In the myth of Prometheus, he stole fire from the gods and delivered this new technology to humanity – who went on to benefit and suffer from the technology of fire through cooking and warfare respectively.  As punishment for this technology innovation, Prometheus was punished with being chained to a rock for eternity and, each day, a bird would come and pick out his liver.

Similarly, we find even resistance to the technology of writing with Socrates famously lambesting the widespread adoption of the written word as an ineffective means of communicating knowledge and Plato worried that we would cease to exercise memory if we rely only on “external marks” (words on paper).

The history of technology adoption is the history of human resistance to the technology with certain patterns reoccuring with each innovation.  As this article explains, even something as seemingly innocuous as the bicycle generated extreme levels of moral and medical panic at the time it was introduced: almost all of this panic directed at the phenomena of ladies riding bicycles.  As the article explains, doctors warned against something called "bicycle face": where the combination of wind on the face and facial strain from the exertion of pushing peddles would permanently distort a woman's face.

Similarily, when the typewriter was introduced, it was characterised as a "cold-blood, mechanical" production that would kill any notions of romance.  The below is an extract from a newspaper in the early 20th warning against this technology.  As we think about the debates as to whether text generated by AI, such as poetry or literature, can be considered "real" or "art", we can see echoes of the debates of the 1900s where typed words were considered "cold-blooded" and it was written that a woman who would accept a typed love letter "will put up with anything".  

As teachers and school officials lament the effects of artificial intelligence, such as GPT, on essay writing and education, it is also useful to remember that, as recently, as the 1980s, teachers were talking about the need to "fight computers".  The below article warns against the "growing invasion of computers in the classroom" that, if not stopped, would make everyone illiterate by the 1990s.

Similarly, the invention of the calculator was met by a backlash in the 1980s with the Chicago Tribune below arguing that, having stood the test of time, the slide rule was a far superior instrument to the calculator:

In 2006, the same drama we see today about cheating with artificiual intelligence was playing out with respect to the proliferation of mobile phones and other technologies n the pages of the New York Times as such "high tech gadgets" were making it easy to cheat and education, as we know it, is under threat.  

The same pattern repeats.  Most all technology, particularly disruptive technologies, tend to be accompanied by forms of moral panic.  If we distill these concerns down to their essence, it seems that it is often not about the technology but about the effects that the technologies have on existing social, economic or political structures.  The animus against women riding bicycles, for example, wasn't about the bicycle but was about the fact that, with increased mobility, came increased independence; the animus against other technologies that appeared over the last century was often driven by parts of the labor force who would be rendered redundant; and the animus about AI today might likewise be less about AI itself and more about how it has the potential to disrupt, displace, or drive restructuring certain parts of the society.

Just as reading about typewriter-driven moral panics, "bicycle face", the impact of electric street lights on proliferating insomnia, the UK's Locomotive Acts requiring all trains to be led by a man waving a red flag, the need to maintain slide rules, or keep computers out of schools seem ridiculous today; we will certainly look back, at some point, on today's moral panics about AI in a similar way.