AI Apocalypse is a Myth: What You Should Really Be Worried About
Fact & Fiction in The Fear Of Artificial Intelligence
Did you know that an AI recently discovered 40.000 new possible chemical weapons in just about six hours?
Or that an AI-powered clone’n’slaughter facility is being built while you are reading this?
Both are actual threats enabled by AI technology. But instead of worrying about it, people rant about ChatGPT and AI-Art and speculate about a malicious, omnipotent AI overlord. What nonsense.
Humans are very good at being malicious overlords themselves. Of course, with the appropriate technology, human cruelty and stupidity reach new heights. If you dare, read on about the AI stuff we should really worry about.
The Myth Of The AI Overlord
“AI will gain consciousness and enslave humanity” — this is one of those absurd claims that circulate whenever Twitter gets excited about an advance in AI tech. As if all you have to do is rig up a bunch of microchips and sensors, and voilá, you have something that happens to have consciousness and an immediate appetite for a killing spree. What a coincidence!
Enslavement Like The Pros
For a moment let’s assume we reach this point: A conscious AI rises up among us. Why the hell then are humans so fixated on being enslaved? Is it possible that this fear comes from deep within us? From the place that knows that human history is actually the history of constant attempts at enslavement, be it physical, mental, or social? In the presence of AI technology, we are once again committing the old mistake of the ape horde that has been shaped by millennia of religious indoctrination: We see ourselves as the crown of creation and therefore assume that artificial intelligence must be like us, the sacred species, because how could anything not bear at least some resemblance to our magnificent species, even if it is only the desire to enslave everything.
The Real Dangers Of Artificial Intelligence
Instead of getting caught up in these absurd fears, let’s be afraid of the amazingly real dangers that AI is bringing as you read this.
For instance:
- an AI just came up with 40.000 new molecules than can be used for chemical warfare
- an AI-powered factory to industrialize cloning and slaughtering is being built
- the AI-assisted manipulation of media content has led to an unprecedented success rate of fake news campaigns
- training biases could lead to upscaling and solidifying of social injustice
Weapon research
All it took an AI to come up with 40.000 new possible chemical weapons was a few lines of code and a weekend’s time. Researchers only had to change their methodology from weeding out to seeking out toxicity. And it would be shockingly easy for anyone with a bit of coding experience to replicate this experiment. Although creating the actual molecules would be a far more difficult task, the researchers were reluctant to publish this finding. They did so, in the end, explicitly as a warning:
“[…] if you were to search for toxicity datasets, there’s a large number of open-source tox datasets. So if you just combine those two things, and then you know how to code and build machine learning models — all that requires really is an internet connection and a computer — then, you could easily replicate what we did.” — Fabio Urbina, lead author of the paper
AI-powered Automated Cloning (and Slaughtering)
Mankind has a great hunger for dead meat. To reduce the country’s dependence on imported pig carcasses, China has started cloning pigs entirely with AI-controlled robots and is working to bring this to industrial standard. This means mass cloning and slaughtering in one hell of a place. That alone is a sad climax to the breathtaking stupidity and cruelty of our species, but there is probably even more to it. You know what they say: if you can clone one animal, you can clone another. The question is not if, but when cloning humans will also be industrialized (whether in the form of space-faring production modules for the noble purpose of space exploration or as cannon fodder for the more earthly purpose of waging war).
“Our AI-powered system can calculate the strain within a cell and direct the robot to use minimal force to complete the cloning process, which reduces the cell damage caused by human hands.” — Liu Yaowei, who contributed to the development of the system and process.
Misinformation
Divide and conquer has become much easier now: With recent advances in AI-generated images, video, text, and sound, Deep Fake Media has reached a whole new level of perfection. It is already possible to convincingly fake anything you want. It is expected that this will be widely used to manipulate people, control communities and manipulate society.
Training Data Bias
Imagine an AI system used to find the best applicants for a job. A bias in the training data could then result in some groups being favored over others. This would be similar to the traditional under-representation of women, except that AI-assisted under-representation would probably reach a whole new level of scrutiny (check out Auntie Horst to see how sexist Midjourney can be). In criminal justice, for example, if the training data contains demographic information that correlates with criminal behavior, the AI system could then associate race or poverty with a higher risk of recidivism. Like a racist society, but automated.
Are you worried now? Great, because these are real dangers in the real world, not some sci-fi-inspired brain fart predicting malicious artificial consciousness. So, if you want to shit your pants, please do so for the right reasons. However, if you want to feel a bit more optimistic again, follow my lead and read the more amusing part of this crazy new tech here, here, and here.