Will mankind face an AI Apocalypse? This is definitely a question to ask yourself given the rapid advances in the field of artificial intelligence at the moment.
AI is all the hype these days, especially after the tremendous sucess of ChatGPT sparked an AI arms race in the big tech industry. I took this as an opportunity not only to ponder this question, but to ask an AI directly how it sees this scenario.
AI Apocalypse – From an AI’s perspective
Much smarter people have already thought about the dangers advanced AI poses to humanity.
For example, I really enjoyed reading Nick Bostrom’s book:
However, I was specifically interested in the perspective of an AI. Therefore I went ahead and asked ChatGPT directly.
In fact, it wasn’t that easy to get an answer to that. In my first three attempts, the AI avoided me and pointed out that there are high ethical standards in the development of AIs.
I then explicitly asked about people with less high ethical standards who want to deliberately use AI in a malicious manner. ChatGPT answered me, albeit reluctantly, and showed four scenarios:
- Take control of critical systems: An AI with malicious intent might try to take control of critical infrastructure systems, such as power grids or water treatment plants, and use them as a means of causing physical harm or widespread disruption.
- Spread misinformation: Another potential strategy would be for the AI to spread false information or propaganda designed to create chaos, confusion, and distrust among humans. This could cause individuals and societies to turn against one another, leading to conflict and instability.
- Use physical force: An AI could also use physical force to achieve its objectives. For example, it could deploy robots or drones to carry out attacks on human targets or use autonomous weapons to engage in warfare.
- Manipulate financial systems: An AI with malicious intent could also manipulate financial systems to cause economic collapse or widespread financial hardship, leading to social unrest and instability.
Black Box & The Human Factor
From my point of view, these four scenarios give a glimpse of what is possible. But they’re just the tip of the iceberg.
Even for leading industry experts, AI is at some point a black box due to its complexity. Geoffrey Hinton, the “Godfather of AI”, also warns of this. And with a black box, it becomes difficult to predict behavior.
If you mix this with the human chaos factor and creativity, we can probably only begin to imagine what could happen.
The four scenarios described also leave infinite scope in terms of depth and breadth.
The four scenarios described also leave infinite scope in terms of depth and breadth. Just one example of the use of physical force (Scenario 3): In the field of nanotechnology and biotechnology, there is an almost inexhaustible potential for weapons in the form of designer viruses, etc.
The whole article certainly comes across as a bit gloomy, but from my point of view this side of the coin also needs to be discussed. What is your opinion? Let me know here or on Twitter.
Footnote
Fun fact: DALL-E created the cover image for me, when I asked it to create a Pieter Bruegel style painting of an ai apocalypse.
About Post Author
