An Inconvenient Doom

An upcoming documentary about AI existential risk and the societal and political discussion about it, directed by @NikiDrozdowski

[An Inconvenient Doom @An_Inconv_Doom]

AI existential risk is hard to explain – but we did it anyway!

Watch the two minute teaser here, and then dive in to the full explanation video of AI existential risk in laypeople's terms – complete with 45 links to articles, research papers and other videos.

This topic concerns literally everybody. So a firm understanding of the basics is the first step towards a fact-based discussion of risks.

Link in 🧵

[An Inconvenient Doom @An_Inconv_Doom]

This is the paper in question: https://t.co/s5iwCRfv9z

[Dan Hendrycks @hendrycks]

We’ve found as AIs get smarter, they develop their own coherent value systems.

For example they value lives in Pakistan > India > China > US

These are not just random biases, but internally consistent values that shape their behavior, with many implications for AI alignment. 🧵

[An Inconvenient Doom @An_Inconv_Doom]

Congratulations to @Malala! According to this ACTUAL research paper you're the only person to survive the AI apocalypse, because the AI values your life more than its own. @elonmusk and @realDonaldTrump on the other hand should seriously consider regulations if they want to live.

[An Inconvenient Doom @An_Inconv_Doom]

@NikiDrozdowski @KKumar_ai_plans @maxwinga Like so: https://t.co/M3P8jQVhwx

[An Inconvenient Doom @An_Inconv_Doom]

@FLI_org Thank you so much!

[An Inconvenient Doom @An_Inconv_Doom]

@burnt_jester @ciphergoth Well, the 31% also have no evidence. This is an ongoing debate. If you’ve watched the video @NikiDrozdowski says in the end, he hopes that he is wrong on this, as do the scientists. But as @Yoshua_Bengio wrote: He has yet to hear a valid argument against the risk.

[An Inconvenient Doom @An_Inconv_Doom]

AI existential risk is hard to explain – but we did it anyway!

Watch the two minute teaser here, and then dive in to the full explanation video of AI existential risk in laypeople's terms – complete with 45 links to articles, research papers and other videos.

This topic concerns literally everybody. So a firm understanding of the basics is the first step towards a fact-based discussion of risks.

Link in 🧵

[An Inconvenient Doom @An_Inconv_Doom]

@burnt_jester @ciphergoth Where is the cutoff? Is climate change not real, because 0,something % of scientists doubt it? Do we need to take THEM seriously and question everything? If the survival of our species is at stake we should indeed play it safe. This is what the documentary will explore.

[An Inconvenient Doom @An_Inconv_Doom]

@burnt_jester @ciphergoth Sorry, I meant experimental proof. But in any case: I’ll stick to the worried majority of AI researchers with this one instead of personal incredulity - just as a precaution. Everything could turn out to be fine in the end - but until then we should take them seriously.

[An Inconvenient Doom @An_Inconv_Doom]

@burnt_jester @ciphergoth So, the Hawking effect, string theory and back in the days relativity should have just been disregarded because there was/is no formal proof? “Deceptive alignment” has been theorized about more than a decade ago and was just observed with OpenAI’ o1 model.

[An Inconvenient Doom @An_Inconv_Doom]

@burnt_jester @ciphergoth True. But the core issue is not an intuition. 69% of AI researchers believe the chance of human extinction from AI is non-zero. And 96% think, the alignment problem is real.

[An Inconvenient Doom @An_Inconv_Doom]

We know the competition is long over, but can we at least get a consolation retweet? @FLI_org

[An Inconvenient Doom @An_Inconv_Doom]

AI existential risk is hard to explain – but we did it anyway!

Watch the two minute teaser here, and then dive in to the full explanation video of AI existential risk in laypeople's terms – complete with 45 links to articles, research papers and other videos.

This topic concerns literally everybody. So a firm understanding of the basics is the first step towards a fact-based discussion of risks.

Link in 🧵

[An Inconvenient Doom @An_Inconv_Doom]

@judge_sneed Thanks, that’s awesome! I think spreading the word helps the most for now.

[An Inconvenient Doom @An_Inconv_Doom]

@NikiDrozdowski @atheorist Which is this one: https://t.co/M3P8jQUJGZ

[An Inconvenient Doom @An_Inconv_Doom]

@skeptycali @ESYudkowsky Here is a 101 video for laypeople: https://t.co/OH1y4kIYUk

[An Inconvenient Doom @An_Inconv_Doom]

AI existential risk is hard to explain – but we did it anyway!

Watch the two minute teaser here, and then dive in to the full explanation video of AI existential risk in laypeople's terms – complete with 45 links to articles, research papers and other videos.

This topic concerns literally everybody. So a firm understanding of the basics is the first step towards a fact-based discussion of risks.

Link in 🧵

[An Inconvenient Doom @An_Inconv_Doom]

@SocraGPT @ESYudkowsky Or if you want an even more general overview: https://t.co/OH1y4kIYUk

[An Inconvenient Doom @An_Inconv_Doom]

AI existential risk is hard to explain – but we did it anyway!

Watch the two minute teaser here, and then dive in to the full explanation video of AI existential risk in laypeople's terms – complete with 45 links to articles, research papers and other videos.

This topic concerns literally everybody. So a firm understanding of the basics is the first step towards a fact-based discussion of risks.

Link in 🧵

[An Inconvenient Doom @An_Inconv_Doom]

@NikiDrozdowski @ESYudkowsky And maybe send him this first: https://t.co/M3P8jQVhwx

[An Inconvenient Doom @An_Inconv_Doom]

@danfaggella Absolutely. There is no second place in the race to AGI. And everybody calling AI a bubble because the labs lose money on the chatbots needs to understand, that today's LLMs are merely a byproduct. AGI is the end that justifes all means for them.

[An Inconvenient Doom @An_Inconv_Doom]

https://t.co/M3P8jQUJGZ

And if you want to know more about the upcoming documentary visit https://t.co/IQKLpTd9H1