On March 31, I made a post about how I think the AI Safety community should try hard to keep the momentum going by seeking out as much press coverage as it can, since keeping media attention is really hard but the reward can be really large. The following day, my post proceeded to get hidden under a bunch of April Fools day posts. Great irony.
I think this point is extremely important and I'm scared that the AI Safety Community will not take full advantage of the present moment. So I've decided to write a longer post, both to bump the discussion back up and to elaborate on my thoughts.
Media coverage of AI Safety is, in my mind, critical in the AI Safety mission. I have two reasons for thinking this.
The first is that we just need more people aware of AI Safety. Right now it's a fairly niche issue, both because AI as a whole hasn't gotten as much coverage as it deserves and because most people who have seen ChatGPT don't know anything about AI risk. You can't tackle an issue if nobody knows that it exists.
The second reason relies on a simple fact of human psychology: the more people hear about AI Safety, the more seriously people will take the issue. This seems to be true even if the coverage is purporting to debunk the issue (which as I will discuss later I think will be fairly rare) - a phenomenon called the illusory truth effect. I also think this effect will be especially strong for AI Safety. Right now, in EA-adjacent circles, the argument over AI Safety is mostly a war of vibes. There is very little object-level discussion - it's all just "these people are relying way too much on their obsession with tech/rationality" or "oh my god these really smart people think the world could end within my lifetime". The way we (AI Safety) win this war of vibes, which will hopefully bleed out beyond the EA-adjacent sphere, is just by giving people more exposure to our side.
(Personally I have been through this exact process, being on the skeptical side at first before gradually getting convinced simply by hearing respectable people concerned about it for rational reasons. It's really powerful!)
Who is our target audience for media coverage? In the previous post, I identified three groups:
- Tech investors/philanthropists and potential future AI Safety researchers. The more these people take AI Risk seriously, the more funding there will be for new / expanded research groups and the more researchers will choose to go into AI Safety.
- AI Capabilities people. Right now, people deploying AI capabilities - and even some of the people building them - have no idea of the risks involved. This has lead to dangerous actions like people giving ChatGPT access to Python's exec function and Microsoft researchers writing "Equipping LLMs with agency and intrinsic motivation is a fascinating and important direction for future work" in their paper. AI capabilities people taking AI Safety seriously will lead to fewer of these dangerous actions.
- Political actors. Right now AI regulation is virtually non-existent and we need this to change. Even if you think regulation does nothing good but slow progress down, that would actually be remarkable progress in this case. Political types are also the most likely to read press coverage.
Note that press coverage is worth it even if few people from these three groups directly see it. Information and attitudes naturally flow throughout a society, which means that these three groups will get more exposure to the issue even without reading the relevant articles themselves. We just have to get the word out.
Why Maintaining Media Coverage Will Take A Lot Of Effort
The media cycle is brutal.
You work really hard to get an article to be written about your cause only for it to get stripped from the front page days later. Even the biggest news stories only last for a median of seven days.
The best way - maybe the only way - to keep AI Safety in the news is just to keep seeking out coverage. As I wrote in the original post:
AI Safety communicators should be going on any news outlet that will have them. Interviews, debates, short segments on cable news, whatever. . . This was notably Pete Buttigieg's strategy in the 2020 Democratic Primary (and still is with his constant Fox News cameos), which led to this small-town mayor becoming a household name and the US Secretary of Transportation.
We should never assume that we have exhausted all options with reaching out to press. Even actions like going on the same program multiple times probably aren't as effective as going on different programs but are still valuable. I admit I'm not exactly sure how one gets interviews and the like, but I'm assuming it's a mix of reaching out to anyone that might seem interested in having you and just saying yes to anyone who wants you on. It's all about quantity.
This is the most controversial of these three points and the one I am least confident in. There is, to some extent, a tradeoff between how clear you make your message and how much you seek out media coverage. This is an optimization problem, and so the answer is obviously not to totally disregard message clarity. That being said, I think we should strongly lean toward the side of chasing media coverage.
In the original post, I alluded to two different factors that might cause someone to turn down media coverage in favor of maintaining a clear message:
a) They don't feel confident enough in their ability to choose words carefully and tell the technical details precisely. To elaborate, I think the people reaching out to reporters should be those knowledgeable in AI Safety and not just anyone vaguely in EA. I do not think this person needs to be highly trained at dealing with the press (though some training would probably be nice) or totally caught up on the latest alignment research. Obviously if the interview is explicitly technical we should have a more technically-knowledgeable person on it.
b) They are afraid of interacting with an antagonistic reporter. To elaborate, I don't think communicators should reach out to reporters who have already done an interview where they treated AI Safety as a big joke or accused it of being a secret corrupt mission. I do think that ~every reporter should be given the benefit of the doubt and assumed not antagonistic, even if we disagree with them on most things and maybe even if they have said bad things about EA in the past.
I think these two factors seem much scarier than they are actually harmful, and I hope they don't cause people to shy away from media coverage. A few reasons for this optimism:
One reason, specifically for part (a), is that the public is starting from a place of ~complete ignorance. Anyone reading about AI Safety for the first time is not going to totally absorb the details of the problem. They won't notice if you e.g. inaccurately describe an alignment approach - they probably won't remember much that you say beyond "AI could kill us all, like seriously". And honestly, this is the most important part anyway. A tech person interested in learning the technical details of the problem will seek out the better coverage and find one of the excellent explainers that already exist. A policymaker wanting to regulate this will reach out to experts. You as a communicator just have to spread the message.
Another reason is that reporters and readers alike won't be all that antagonistic. EAs are notably huge contrarians and will dig in to every single technical detail to evaluate the validity of an argument, probably updating against the point if it is not argued well enough. Most people are not like that, particularly when dealing with technical issues where they are aware that they know far less than the person presenting the problem. Also because AI safety is a technical issue, you don’t get this knee-jerk antagonism that happens when people’s ideology is being challenged (ie when you tell people they should be donating to your cause instead of theirs). My guess is that <25% of pieces will be antagonistic, and that when reading a non-antagonistic piece <25% of readers will react antagonistically. We don't need a security mindset here.
The final reason, perhaps the biggest point against any of these fears, is that any one individual news story is not going to have much impact. In the worst case scenario, someone's first exposure to AI Safety will be an antagonistic reporter treating it as a big joke. Their second and third and fourth exposures will likely be of reporters taking it seriously, however. It's not a huge deal if we have a couple of bad interviews floating around the sphere - so long as the coverage is broadly serious and correct, it will be worth it.
We have just been handed a golden opportunity with this FLI letter. Let's not mess this up.
I think when trying to get people concerned about "AI safety," there are a number of "nearby messages" or misunderstandings that are easy for other people to hear, and are potentially very net negative to spread. For example, if you are trying to communicate "AGI systems will one day be very powerful and dangerous, they are by default controllable, therefore we should slow down and try to develop alignment techniques first" people might instead hear:
I'm not saying these issues in communication are insurmountable. But I think they're pretty serious, and have arguably historically led to a lot of x-risk. So I think we should not be cavalier in communications about AI safety. The strongest counterargument I'm aware of is something like "This time is different" with the idea that AI progress is already so much in the news that capabilities externalities are minimal. But I think this is not currently the world we live in, and I will be pretty worried about spreading messages that "AI is powerful" or even "AI is powerful and dangerous" to national security folks, or political actors more generally.
I think the best case scenario is convincing people who already believe "AGI is powerful" that AGI is dangerous, rather than messaging on both issues at once. To that end, I think prematurely chasing AI safety messages is bad.
Not to be rude but this seems like a lot of worrying about nothing. "AI is powerful and uncontrollable and could kill all of humanity, like seriously" is not a complicated message. I'm actually quite scared if AI Safety people are hesitant to communicate because they think the misinterpretation will be as bad as you are saying here; this is a really strong assumption, an untested one at that, and the opportunity cost of not pursuing media coverage is enormous.
The primary purpose of media coverage is to introduce the problem, not to immediately push for the solution. I stated ways that different actors taking the problem more seriously would lead to progress; I'm not sure that a delay is actually the main impact. On this last point, note that (as I expected when it was first released) the main effect of the FLI letter is that a lot more people have heard of AI Safety and people who have heard of it are taking it more seriously (the latter based largely on Twitter observations), not that a delay is actually being considered.
I don't actually know where you're getting "these issues in communication...historically have led to a lot of x-risk" from. There was no large public discussion about nuclear weapons before initial use (and afterwards we settled into the most reasonable approach there was for preventing nuclear war, namely MAD) or gain-of-function research. The track record of "tell people about problems and they become more concerned about these problems", on the other hand, is very good.
(also: premature??? really???)
To first order, the problem isn't that the message is complicated. "Bioterrorism might kill you, here are specific viruses that they can use, we should stop that." is also not a complicated message, but it'll be a bad idea to indiscriminately spread that message as well.
Well there was DeepMind, and then OpenAI, and then Anthropic.
I don't view this as a crux. I weakly think additional attention is a cost, not a benefit.
I meant in AI. Also I feel like this might be the crux here. I currently think that past communications (like early Yudkowsky and Superintelligence) have done a lot of harm (though there might have been nontrivial upsides as well). If you don't believe this you should be more optimistic about indiscriminate AI safety comms than I am, though maybe not to quite the same extent as the OP.
Tbh in contrast with the three target groups you mentioned, I feel more generally optimistic about the "public's" involvement. I can definitely see worlds where mass outreach is net positive, though of course this is a sharp departure from past attempts (and failures) in communication.
Ahh, I didn't read it as you talking about the effects of Eliezer's past outreach. I strongly buy "this time is different", and not just because of the salience of AI in tech. The type of media coverage we're getting is very different: the former CEO of Google advocating AI risk and a journalist asking about AI risk in the White House press briefing is just nothing like we've ever seen before. We're reaching different audiences here. The AI landscape is also very different; AI risk arguments are a lot more convincing when we have a very good AI to point to (GPT-4) and when we have facts like "a majority of AI researchers think p(AI killing humanity)>10%".
But even if you believe this time won't be different, I think we need to think critically about which world we would rather live in:
Let's even think about the history here. Early Eliezer advocating for AGI to prevent nanotech from killing all of humanity was probably bad. But I am unconvinced that Eliezer's advocacy from afterwards up until 2015 or whatever was net-negative. My understanding is that though his work led to development of AI capabilities labs, there was nobody at the time working on alignment anyway. This reflex of "AI capabilities research bad" only holds if there is sufficient progress on ensuring AI safety in the meantime.
One last note, on "power". Assuming Eliezer isn't horribly wrong about things, the worlds in which we survive AI are those where AI is widely acknowledged as extremely powerful. We're just not going to make it if policy-makers and/or tech people don't understand what they are dealing with here. Maybe there are reasons to delay this understanding a few years - I personally strongly oppose this - but let's be clear about this.
Anything longer than 8 morphemes is probably not going to survive Twitter or CNN getting their hands on it. I like the original version ("Literally everyone will die") better.
I usually agree with @Linch , but strongly disagree here. I struggle to understand the causal pathways for which misunderstanding or "nearby messages" are going to do more harm than good. I also think the 4 thoughts that were bullet pointed are unlikely misunderstandings. And even if people did hear those, it's good that people have started thinking about it.
More coverage = better, and the accuracy and nuance isn't so important right now.
I will just copy paste from the OP because they put it so well.
"the public is starting from a place of ~complete ignorance. Anyone reading about AI Safety for the first time is not going to totally absorb the details of the problem. They won't notice if you e.g. inaccurately describe an alignment approach - they probably won't remember much that you say beyond "AI could kill us all, like seriously". And honestly, this is the most important part anyway. A tech person interested in learning the technical details of the problem will seek out the better coverage and find one of the excellent explainers that already exist. A policymaker wanting to regulate this will reach out to experts. You as a communicator just have to spread the message.
I've got a friend who often says (kind of jokingly) "We're all going to die" when he talks about AI. It gets people interested, makes them laugh and gets the word out there.
I think AGI research is bad. I think starting AGI companies is bad. I think funding AGI companies is bad. I think working at AGI companies is bad. I think nationalizing and subsidizing AGI companies is probably bad. I think AGI racing is bad. I think hype that causes the above is bad. I think outreach and community building that causes the above is bad.
Also the empirical track record here is pretty bad.
Agree with this 100% "I think AGI research is bad. I think starting AGI companies is bad. I think funding AGI companies is bad. I think working at AGI companies is bad. I think nationalizing and subsidizing AGI companies is probably bad. I think AGI racing is bad."
Thanks, are you arguing that raising AI safety awareness will do more harm than good, through increasing the hype and profile of AI? That's interesting will have to think about it!
What do you mean by "the empirical track?"
Thanks!
The empirical track record is that the top 3 AI research labs (Anthropic, DeepMind, and OpenAI) were all started by people worried that AI would be unsafe, who then went on to design and implement a bunch of unsafe AIs.
100% agree. I'm sometimes confused on this "evidence based" forum as to why this doesn't get the front page attention and traction.
At a guess, perhaps some people involved in the forum here are friends, or connected to some of the people involved in these orgs and want to avoid confronting this head on? Or maybe they want to keep good relationships with them so they can still influence them to some degree?
Should be "empirical track record." Sorry, fixed.