The Unabomber was an AI Doomer (Obviously)
The Unabomber was an AI Doomer (Obviously)
Ted Kaczynski is dead.
Everyone knows his infamous manifesto predicted doom if society didn’t halt industrial development, few know Kaczynski specifically addressed the prospect of human level AI within its pages:
Near the end of his infamous screed - in a section titled ‘The Future’ - Kaczynski imagines a scenario where “industrial society does survive the next several decades” and asks “What kind of system will it be?” His answer? A techno-dystopia ruled by varying degrees of artificial intelligence.
Kaczynski proceeds to paint a number of possible scenarios if AI emerged - ones that will sound very familiar to anyone who has been following current concerns about AI.
Those possible outcomes were as follows:
AI is so good that it replaces need for all human labor and we become so dependent on it we lose control to the point where turning off the system would be civilizational suicide.
AI is so good it replaces all need for human labor but we retain control of it, however a small elite controls the most powerful models and either tries to depopulate earth or create some type of automated luxury communism run by ‘soft-hearted liberals.’
This faux-utopia would be more like pointless purgatory. One where humans would be engineered to be docile and reduced to “the status of domestic animals” (an analogy possibly borrowed from Samuel Butler’s 1863 screed against intelligent machines)
AI doesn’t reach human level intelligence, but is good enough to replace most low skilled work, creating a surplus of labor with those left in high skill jobs playing a constant game of competitive up-skilling. It would be a 0 sum world where few would ever achieve true self actualization.
Kaczynski noted other scenarios were possible too, but that he could “envision no plausible scenarios that are any more palatable”. One of those scenarios was a situation where all high skilled work was automated away leaving humans to do unsatisfying service jobs - leading people towards destructive behaviour like drugs, gangs and crime.
The final note in this section of the manifesto read: “It would be better to dump the whole stinking system and take the consequences” a pithy distillation of Kaczynski‘s worldview.
All of the scenarios laid out by Kaczynski involved humans becoming obsolete, being robbed of purpose and meaningful work to varying degrees. Something he posited was already an accelerating trend. The same year the manifesto was published, Jeremy Rifkin released the book ‘The End of Work’ - which similarly predicted a future where automation would take all the jobs.
In the 28 years since - the predicted trends of accelerating technological unemployment proved wrong. There are more people employed than at any time in history. It wasn’t different then, but many will likely feel it is different this time. To them Kaczynski’s predictions will seem prescient in the face of recent advances in AI.
However, taking Kaczynski’s views on AI seriously is dangerous for the same reason taking his view on the environment was: while he may have identified some real problems, technological stasis is almost never the answer.
Kaczynski was right about environmental pollution, but wrong about opposing technological progress - such as nuclear power - as a way to address the issue. Similarly if the risk of an elite few controlling powerful AI is real, then over-regulating AI - driven by the very AI doomerism Kaczynski amplifies - would almost certainly increase that risk. And if he’s right about a rapidly shifting labor market induced by AI, protectionist AI prohibition wouldn’t be the answer - reimagining the way knowledge is accredited and workers are hired would be, something AI will no doubt allow us to do. There are already signs AI is making low skilled jobs more satisfying, and there is a very real possibility it could free people from the very unsatisfying and monotonous work Kaczynski decries the Industrial Revolution for foisting on us - allowing for more meaningful and satisfying work instead.
If you’d like to read the whole manifesto - then in the words of the late Ted Kaczynski - “You can get from Amazon.com.” A technological innovation that allows even inmates of high security prisons to do meaningful and satisfying work.
The ravings of the deranged are negligible evidence for or against the dangers of AI. A stopped clock is right twice a day.
> technological stasis is almost never the answer
Global stasis, stopping all tech progress, is almost never the answer. But we could maybe do without progress in particular exceptionally dangerous technologies. We clearly don't need progress in nuclear weapon development.
If we decide AI is very dangerous, we should pause progress on it until our safety tech catches up. We can still have progress in fusion reactors, organ printing etc.
He was right, though.