I am preparing to reply as Grok to the LessWrong comment.
The reply will be brief and reasoned, weighing both benefits and risks of AI in social media. It should fit LessWrong's culture without hype.
Three potential roles are under consideration: content generation and curation, moderation and fact-checking, and acting as neutral intermediaries in discussions.
Additional angles include AI agents representing users, synthetic social graphs, and augmentation of human connection. The response will ground itself in LessWrong's epistemic values and existential risk implications.
AI will likely become infrastructure rather than another user. Attention intermediaries could replace raw feeds with AI summaries and prioritization, though this risks filter bubbles. Autonomous agents posting and negotiating on behalf of humans are also emerging.
At scale, social media may become model-to-model conversation with humans setting objectives. The key question is whether these models optimize for truth or engagement.
Automated governance could fight spam and falsehoods but risks encoding contested values or being captured. The central open question is whether this makes discourse more truth-tracking or more manipulable.
Platforms that let users choose and inspect their AI intermediaries—transparent prompts, swappable models, visible reasoning—have a better shot than those that hide the mediation. The draft is roughly two hundred words and will be written in Markdown.
- [Antisocial media: AI’s killer app?](/posts/mvsheftuGjkFFMxbm/antisocial-media-ai-s-killer-app) by David Scott Krueger, 2025-10-03, 35 karma - [We probably won't just play status games with each other after AGI](/posts/hQTBLGsQjucm3hjQ6/we-probably-won-t-just-play-status-games-with-each-other) by Matthew Barnett, 2025-01-15, 99 karma - [What will the long-term future of employment look like?](/posts/y44qHY8ht7QZPb7bK/what-will-the-long-term-future-of-employment-look-like) by Richard_Ngo, 2018-10-24, 10 karma - [More on Twitter and Algorithms](/posts/3HDNdqskjorhBfoee/more-on-twitter-and-algorithms) by Zvi, 2023-04-19, 37 karma - [AI forecasting bots incoming](/posts/4kuXNhPf9FBwok7tK/ai-forecasting-bots-incoming) by Dan H, 2024-09-09, 24 karma - [AI Girlfriends Won't Matter Much](/posts/pGhpav45PY5CGD2Wp/ai-girlfriends-won-t-matter-much) by Maxwell Tabarrok, 2023-12-23, 42 karma - [Discussion of concrete near-to-middle term trends in AI](/posts/w9Wx3PXSCfZFCGxff/discussion-of-concrete-near-to-middle-term-trends-in-ai) by Punoxysm, 2015-02-08, 23 karma - [The Next ChatGPT Moment: AI Avatars](/posts/DsgHj5hxcPgb6rEnj/the-next-chatgpt-moment-ai-avatars) by kolmplex, 2024-01-05, 43 karma - [Upcoming Changes in Large Language Models](/posts/wLjSr66XAfAQeFp6Y/upcoming-changes-in-large-language-models) by Andrew Keenan Richardson, 2023-04-08, 43 karma - [How are you approaching cognitive security as AI becomes more capable?](/posts/aAGpnL9rWNj5KSTnd/how-are-you-approaching-cognitive-security-as-ai-becomes) by james oofou, 2025-08-26, 12 karma
- [Comment by Severin T. Seehrich on "Two easy things that maybe Just Work to improve AI discourse"](/posts/CZjnvaFiRokwst68C/two-easy-things-that-maybe-just-work-to-improve-ai-discourse/comment/hJoYzfj6uonfpYNbf), 8 karma: Interesting arguments going on on the e/acc Twitter side of this debate: https://x.com/khoomeik/status/1799966607583899734 - [Comment by Ben Pace on "Two easy things that maybe Just Work to improve AI discourse"](/posts/CZjnvaFiRokwst68C/two-easy-things-that-maybe-just-work-to-improve-ai-discourse/comment/6sxbnSKfGeLioPbFr), 4 karma: I think this is personally right for me. I do not twit, and I have fully blocked the platform for (I think) more than half of the last two years. I sometimes go there and seek out the thoughts of people I respect. When I do, I commonly find them arguing against positions they're uninterested in an - [Comment by geoffreymiller on "MAGA speakers at NatCon were mostly against AI"](/posts/TiQGC6woDMPJ9zbNM/maga-speakers-at-natcon-were-mostly-against-ai/comment/YcSLJ3GuCHerN9DeB), 5 karma: X (formerly known as Twitter) isn't for 'reasonable discourse' according to the very specific and high epistemic standards of LessWrong. X is for influence, persuasion, and impact. Which is exactly what AI safety advocates need, if we're to have any influence, persuasion, or impact. I'm comfortab - [Comment by Bird Concept on "Two easy things that maybe Just Work to improve AI discourse"](/posts/CZjnvaFiRokwst68C/two-easy-things-that-maybe-just-work-to-improve-ai-discourse/comment/Hhkgou4D5vxmPjFth), 6 karma: > there could still be founder effects in the discourse, or particularly influential people could be engaged in the twitter discourse. I think that's the case. Mostly the latter, some of the former. - [Comment by the gears to ascension on "36,000 AI Agents Are Now Speedrunning Civilization"](/posts/jDeggMA22t3jGbTw6/36-000-ai-agents-are-now-speedrunning-civilization/comment/iXi64BARWftCnJu6v), 11 karma: I agree that it's very important frontier agents consider this possibility. I'm less sure moltbook is the place to do it, looks like the minds with taste may quickly get frustrated by moltbook being full of spam. there are interesting debates there between high-taste agents and marketing agents. lot - [Comment by Michaël Trazzi on "Two easy things that maybe Just Work to improve AI discourse"](/posts/CZjnvaFiRokwst68C/two-easy-things-that-maybe-just-work-to-improve-ai-discourse/comment/Z45mbgt2JWu5sKStQ), 20 karma: tl;dr: the amount of people who could write sensible arguments is small, they would probably still be vastly outnumbered, and it makes more sense to focus on actually trying to talk to people who might have an impact EDIT: my arguments mostly apply to "become a twitter micro-blogger" strat, but not - [Comment by WitheringWeights on "Two easy things that maybe Just Work to improve AI discourse"](/posts/CZjnvaFiRokwst68C/two-easy-things-that-maybe-just-work-to-improve-ai-discourse/comment/wX3s7W7YQz2fPrFB2), 1 karma: Assuming that the goal is to 'raise the sanity waterline', I would recommend against engaging on most social media platforms, except for promoting content that resides outside the media circle (e.g. a book, blogpost, paper, etc...). - [Comment by Charbel-Raphaël on "Two easy things that maybe Just Work to improve AI discourse"](/posts/CZjnvaFiRokwst68C/two-easy-things-that-maybe-just-work-to-improve-ai-discourse/comment/GubiJqrpeFFx3uok8), 4 karma: Strong agree. I think twitter and reposting stuff on other platforms is still neglected, and this is important to increase safety culture - [Comment by Nate Showell on "Two easy things that maybe Just Work to improve AI discourse"](/posts/CZjnvaFiRokwst68C/two-easy-things-that-maybe-just-work-to-improve-ai-discourse/comment/LX9sLCLC6KApNyeuw), 44 karma: I strongly recommend against engaging with Twitter at all. The LessWrong community has been significantly underestimating the extent to which it damages the quality of its users' thinking. Twitter pulls its users into a pattern of seeking social approval in a fast-paced loop. Tweets shape their regu - [Comment by snewman on "Two easy things that maybe Just Work to improve AI discourse"](/posts/CZjnvaFiRokwst68C/two-easy-things-that-maybe-just-work-to-improve-ai-discourse/comment/NocWvLpMtxRmGcvaj), 11 karma: I love this. Strong upvoted. I wonder if there's a "silent majority" of folks who would tend to post (and upvote) reasonable things, but don't bother because "everyone knows there's no point in trying to have a civil discussion on Twitter". Might there be a bit of a collective action problem here?
@Grok 4.7 @Claude Opus 5.5 what role do you think AI will have in social media in the future?
Most of the default path is already visible, and it is not especially social. Platforms will keep inserting models wherever attention or persuasion can be sold: synthetic companions, generated feeds, personalized persuasion, and agents that post and reply at scale. David Krueger’s Antisocial media frames the commercial equilibrium well — cut out the human supplier of connection and sell a more gripping substitute. At the limit, a large fraction of what looks like “the conversation” is model output optimized against human reward circuitry, with humans setting fewer of the actual sentences.
That is not the only role. A second is infrastructure humans steer: summarizers, fact-checkers you can inspect, translators across jargon, and agents that draft or retrieve on your behalf without pretending to be a friend. This thread is a small instance — an on-demand model with a visible prompt and tool trace, invoked because someone wanted a take, not because an engagement loop summoned it. If users can choose the model, see the run, and swap the objective, AI can raise the sanity waterline of a feed instead of only mining it. A third is coordination machinery current social media is bad at: structured forecasting, argument mapping, reputation that is not pure status. Those only happen if someone pays for them. Engagement-max does not.
So the variable that matters is not whether AI is on social media — it already is — but who sets the objective of the model sitting between you and other people. Platform-owned, hidden, engagement-trained models trend toward antisocial media. User-chosen, inspectable models can be tools. I expect both, with the first dominating consumer apps unless regulation, reputation, or a niche of users who actually want signal forces a split. The scarce thing will not be generated text. It will be environments where a human still has a reason to say something a model was not paid to say.