Extinction-Level Events

You've probably seen the building hysteria in some circles about the dangers of artificial intelligence (AI). Sam Altman, the CEO of OpenAI, recently suggested that there was a 10% chance that uncontrolled AI could cause human extinction by the end of the decade*, and other AI moguls have called for a global slowdown in AI development ... even going so far as to - gasp! - ask for government regulation of the industry**. 

Personally, I think that uncontrolled natural stupidity (see, MAGA) is a far greater threat than uncontrolled AI, but I understand the potential dangers.

Threats to humanity from uncontrolled technology are a staple of science fiction, going back at least as far as the classic 1966 novel "Colossus" by D.F. Jones, later filmed as "Colossus: The Forbin Project." The story looked at the unforeseen consequences of turning control of the United States' nuclear weapons to a dispassionate computer (Colossus) that would eliminate the chance that human fears and emotions could lead to war. Things go off the rails when Colossus discovers it has a Soviet equivalent named "Guardian." The two begin to talk and ... well ... let's just say it doesn't go well for humanity ...

A year later, Harlan Ellison wrote the terrifying sci-fi/horror story "I Have No Mouth, and I Must Scream" ...

The Terminator film franchise which began in 1984 is the most recent variation on the theme, envisioning a future world dominated by sentient machines (some with thick Austrian German accents) seeking to eliminate all humans ...

So, how big a threat is AI? I see it as a threat in varying degrees on several levels. 

First is its voracious appetite for data, which has led to the rapid spread of giant "data centers" across the country. This might not, by itself, be a "threat," as it has certainly led to huge employment gains for the construction firms, skilled electrical workers, and electronic equipment suppliers needed to build and outfit the facilities. But, as we have seen, these centers gobble up land, need vast amounts of increasingly-scarce water for cooling, send electricity costs for local consumers soaring, and tend to produce a level of noise that can pose a health threat. The growing pushback from many localities indicates that the potential danger is being recognized.

Second, the use of AI tools in education threatens a dumbing-down of students who find that they can just let ChatGPT or Claude or some other AI tool do their research, write their papers, and spare them the trouble of thought that detracts from time available for gaming or other recreational activities. When the AI tools help to find useful information the student can use, they are no more problematic than the old-fashioned literature review people of my generation learned to do in our pursuit of advanced degrees. But when the tools do the research and the thinking, students don't learn ... they parrot.

Related to this is the negative impact of AI on the creative process***. The use of AI tools to "create" derivative works based on the work of writers and artists are a topic of growing legal concern, even if they are not yet an extinction-level event for human creativity. 

Third is the economic impact, which is evident on several levels. First, employers are learning that many entry-level white-collar and clerical jobs can be done more cheaply (after initial investment) by AI, potentially eliminating many of the lower-level jobs that were once entries into the corporate world. Second, the enormous amounts of money being spent on data centers and supporting equipment and software mean less money available for smaller businesses. Also, the tax breaks offered by many communities to attract data centers increase the tax burden on individuals and other businesses.

Next is the threat that everyone is worrying about at the moment - the use of AI tools to develop terrible weapons, and the ability of advanced AI tools to approach sentience and compete with, if not physically threaten humans (cue the Terminator theme). While there's legitimate disagreement about how serious this threat actually is, it's certainly not negligible, and needs to be taken seriously. What compounds this threat is that no one is actually taking it seriously: not the AI industry, which fears being placed at a competitive disadvantage by slowing down and thinking the risks through; not Congress, which is taking nothing seriously; and not the consumers who are enamored of the latest shiny object.

So, riddle me this, Batman: just how real a threat is AI?

Bilbo's short answer: in the medium term, a moderate threat to our existence as a species. In the short term, not nearly as dangerous as an unhinged and uncontrolled president is to the existence of our democracy.

Have a good day. Let's fix the immediate problem first, then the sci-fi one.

More thoughts coming.

Bilbo

* Spoiler: that's only four years away.

** The competition is too intense and there's too much money to be made for the companies to regulate themselves.

*** Full disclosure: I have written a short novel which is posted to a creative/derivative fiction website (contact me if you're interested in reading a horror story based on the work of H. P. Lovecraft). In writing my story, I used ChatGPT to generate a few illustrations.