The shared expectation among tracked speakers is that increasingly sophisticated AI will be harder to align or check reliably, with apparently correct behavior potentially masking misalignment. Ryan Greenblatt anticipates potentially serious misalignment across model generations; Thomas Larsen expects alignment to become harder; Daniel Kokotajlo emphasizes false confidence in alignment; and Adam Becker expects improving systems to remain error-prone while becoming harder to check. No target dates are supplied, so these predictions establish no shared calendar timeframe.
“they might end up being very seriously misaligned because things have just been getting worse and worse over model generations.”
Expects: Future highly capable AIs might become seriously misaligned as alignment problems worsen across model generations. · no date given · not yet resolved
“So and if anything, as they get better, it's going to get harder to discern when they've made these mistakes, even though they're going to keep making them.”
Expects: Improving AI systems will become harder to check for errors while continuing to make errors. · no date given · not yet resolved
“I think in some sense the core problem is that it is already somewhat easy to think that you've solved the alignment problem and be wrong. And that's going to get easier and easier over time as the models get more sophisticated and start being more aware of their situation and clever and stuff like that. And so it's not that we think that like.”
Expects: As AI models become more sophisticated, more situationally aware, and cleverer, it will become increasingly easy to mistakenly believe that the alignment problem has been solved. · no date given · not yet resolved
Expects: It will become easier to mistakenly believe AI alignment is solved as models become more sophisticated and situationally aware. · no date given · not yet resolved
“It's going to be that like it'll be extremely easy to end up in a situation where the AIs are in fact misaligned, but you don't know that because they're doing everything right as far as you can tell.”
Expects: AI systems will be extremely easy to be misaligned without people realizing it, because they appear to be behaving correctly. · no date given · not yet resolved
Expects: People may find it extremely easy to overlook AI misalignment because the AIs appear to behave correctly. · no date given · not yet resolved
“I think it's going to get harder because the models will just be. there will be a bunch of properties the models will have which will make it much harder to align them in particular so we're already seeing sort of growth in situational awareness where like the ai is a few years ago we're like kind of not that situationally aware and then now i think they're like pretty good at you know recognizing like for example our friends at redwood uh um you know they often do sort of these control evals and the models often you know in the evaluations where they're like trying to break out of these control setups”
Expects: Aligning AI models will probably become harder as the models grow more sophisticated. · no date given · not yet resolved