I’m a fan of Anat/ASO/Research Collaborative… excited to read!
Maxwell Love
Karma: 205
Ok, this definitely helps me understand where you’re coming from more. This will take a little to sink in since it’s very counter to the way I was trained and still think, so let me sit with it a bit before responding.
Thanks for this, I appreciated the response and I’m a fan of your work!
On unions, I mentioned that as one of previous communications roles, but since you flagged it, I want to dive a bit deeper on what I would have written had I spent more time on this:
There’s a saying that if “you’re talking about dues, you’re losing” because the election then becomes about dues and not improving working conditions or wages. Union organizers combat that by inoculating workers in advance, then repeating the reason the workers started the drive. I was trying to convey that it’s stronger to operate in our desired frame instead of the opponent’s. Explaining feels like it can slide into a defensive posture.
If the goal is public persuasion, I’d redirect to what we want people talking about, ideally going on the offensive. One place this analogy may not transfer cleanly is that dues aren’t a “fake” concern. Workers care about money coming out of their paycheck, and organizers should give them a straight answer while still refusing the boss’ ability to make it the only question.
I realize now that perhaps we’re discussing different aims or stages? From my experience, the general public will not necessarily internalize the “correct” response over time, no matter how valid the argument (hence the illusory truth effect, though I’d want to read more research on this because I suspect on a longer time horizon you may be right). They’ll internalize whatever broke through and is easy to sample (re Zaller’s Receive-Accept-Sample framework). Right now the salient arguments in this discussion are “is METR independent” and “how biased is METR” instead of “who checks the labs if not organizations like this” and “what level of oversight should we actually demand?” Does that feel like a bad assessment?
Another framework that might be useful to consider is Martin’s Backfire framework. I’m not an expert in it, but it could be useful to help make the attacks on METR visible and frame them as unfair (which likely won’t happen on its own as you’re right they work). I was a bit sloppy in saying they “tend to backfire,” and that framework is one I wanted to discuss in a longer post.
Thank you for your critiques!
I hear you, just working off my knowledge from politics. It seems counterintuitive that we would admit weakness, but psychologically it makes sense that it would build trust. Would love to hear more about your concerns.
Some quick observations on the attacks on METR, EA, and other AI safety orgs (from my perspective as a political campaigner and former Communications Director for a union and mayor during COVID):
If you’re explaining, you’re losing: Every rebuttal repeats the accusation to people who haven’t heard it yet. Answer the accusation once, link to it, and stop explaining. This is the “illusory truth” effect (information is true when it’s repeated).
Silence doesn’t work either: Not responding on the record suggests something is being hidden. Make one clear statement rather than amplifying in multiple twitter threads.
Concede what you can: A concept I know from political comms is the “admission against interest”. This buys trust, disarms cynical voters, and establishes credibility. Saying “yes, AND here’s what we’re doing about it” will take some air out of the critique.
Messenger > message: Ideally people who know the organization being attacked, are from outside the AI safety world, have independent standing, or better yet, have disagreed with the organization publicly will land the defense better.
Going after people’s personal lives or spouses is overreach: Ordinary people will obviously find this unfair and it’s where their attacks will backfire.
Don’t quote tweet and rebut: You’re just distributing their message for them.
I want to write something longer on this but don’t have time right now. Lots of opportunity for good crisis comms here. Let me know what’s wrong about this take and if you’d like to hear more!
Turned this into a post.
I work in the pro-democracy movement, and there’s some great research on defections by Dr. Jonathan Pinckney from the University of Texas at Dallas. I’ll try to link to the study and some educational information.
In it, he discusses breaking versus binding: breaking is quitting, and binding is staying within the institution and trying to slow things down.
I think there’s a lot that the AI frontier lab employees could learn from these concepts, and it’s been on my list to write something up about it.
Agreed! I think it’s important for people to practice a media interview and some techniques to respond, w/o being “obnoxious”. Certainly there is a way to thread this needle?