Between ChatGPT's surprisingly human voice and Meta's AI characters, we may be witnessing the rise of a new consumer internet era featuring synthetic companions
It's been one of those weeks. One after the other, product launches … Maxwell Timothy / MakeUseOf : What Is MetaAI? And Can It Compete With Other Chatbots? Abdullah / Gizchina : WhatsApp Beta Introduces Quick Reply for Multimedia Steven Levy / Wired : Hey Zuck, Get Those Robots out of My Social Feed Kuwait Times : Meta tech chief defends tech titan's AI strategy The Windows Club : WhatsApp voice message not working on Android, iPhone or PC Threads: Peter Constantine / @salmonbottle : I don't think celebrity chatbots are an actual product. They're a tech demo with the celebrities as hooks to get mainstream consumers excited. No one is going to use these more than a handful of times before the novelty wears off, just like how nobody buys a 4k TV to watch closeup shots of iguanas or slow motion dancing with buckets of paint. … Eugene Morgulis / @elmorgulis : I thought it was brilliant to humanize chatbots, since just chatting is what most people will do (I don't think chatbots are actually very useful for most people on a daily basis). But I was weirded out by giving them celeb faces with different names. Chat sports with Tom Brady? Amazing! … Dare Obasanjo / @carnage4life : My feelings are somewhat mixed on generative AI personalities and social media. The focus of social media apps has shifted over time from connecting us with people to connecting us with entertainment. AI personalities in your messaging app like Snapchat, Messenger & WhatsApp now have extends this disconnection from being about connecting with people. … Benedict Evans / @benedictevans : There's a lot of talk of ‘Thin wrappers on ChatGPT’, but we don't call startups that run on AWS ‘thin AWS wrappers’. There's a lot more to product than the stuff you do to/with the data. It seems to me the real question is whether LLMs will generalise much more, much further up the stack, than any previous platform shift. … Bluesky: @manpageman.bsky.social : This sounds terrible. Endlessly upbeat people with no inner life? No, I prefer to make friends outside of tech. — More riskily, this sort of thing risks placing people in stronger echo chambers and stunting their intellectual growth. Fuck that noise. X: @lizweil : @mmitchell_ai @CaseyNewton The extent to which tech doesn't seem to have an intuitive understanding that people care about people will never stop shocking me. Ryan Hickman / @ryanmhickman : Tried two of the Meta AI bots inside WhatsApp today 🧠 They matched their personas well. Most impressive was how FAST the replies were. Very curious about the inference infra they're using 🤔 [image] Neil Turkewitz / @neilturkewitz : @mmitchell_ai @CaseyNewton What Casey sees as opportunities I see as dangers. He writes “The power of giving what were then text messages a warm & kindly voice, I think, should not be underestimated.” I see that & would suggest that the power to exploit insecurities & mislead should not be underestimated. Steven Levy / @stevenlevy : Facebook/Meta was founded on a vision based on connecting people. You would have more contact with those on your “social graph” and the world would benefit. But mingling AI bots into those feeds. as is the current plan, is the total opposite of social. https://www.wired.com/... Ted Haigh / @dr_cocktail : The synthetic social network is coming https://www.theverge.com/... via @Verge People attach personalities to animals, stuffed toys, cartoons - virtually anything humans feel something about. Synthetic AI personas designed for the purpose are shoe-ins with no barriers whatsoever. @mmitchell_ai : I really appreciated this thoughtful piece from @CaseyNewton. “...Feeds that were once defined by the connections they enabled between human beings will have become something else: a partially synthetic social network.” https://www.theverge.com/... James Nisbet / @bandit : I would absolutely use an AI therapist. Therapy for me is having someone to listen, empathise, and remember. AI is perfect for that https://www.theverge.com/... @iethics : “[S]omething changes when you begin speaking with the app in one of its five native #voices, which are... earnest, upbeat, and—by nature of the fact that they are powered by an #LLM—tireless”: https://www.theverge.com/... #ethics #AI #tech #chatbots #business h/t @mmitchell_ai R.D. Pohl / @rdpohl1 : Love how the tech giants and social media networks were trying to shut down bots, until they introduced their own proprietary AI powered ones. https://www.theverge.com/... @iethics : “You can imagine the next steps here. A #bot that gets to know your quirks; remembers your life history; offers you coaching or tutoring or therapy; entertains you in whichever way you prefer”: https://www.theverge.com/... #ethics #law #AI #chatbots #tech #privacy #business Adam Graber / @adamgraber : On the relational benefits of ChatGPT—this is a longstanding tech narrative: it will heal or help people in need. Be assured: It never stays that way. https://www.theverge.com/... [image] Adam Graber / @adamgraber : Besides the similar voice interface features, this article's comparison of OpenAI and Facebook may be portentous. Today's OpenAI may be 2006's Facebook. And I wouldn't be surprised if 2016's Facebook portends 2033's OpenAI. Political fallout included. https://www.theverge.com/... Forums: Rglullis To / Beehaw : The synthetic social network is coming
Context & Ripple Effects
Meta had already positioned AI as infrastructure for improving Reels engagement and ad targeting. Bringing AI personalities into messaging and social products extends that strategy from behind-the-scenes optimization to a user-facing interaction layer.
The story is an early marker in an arc later visible in a bot-populated social-network experiment and more personal companion use. Its importance is less any single character than the prospect that synthetic interaction becomes native to consumer platforms.
First-order effects
- ChatGPT’s more humanlike voice makes conversational AI feel less like text search and more like a persistent interlocutor for users.
- Meta’s AI characters place synthetic personalities inside WhatsApp and its other social products, giving the company a new type of interaction to design, distribute, and moderate.
Second-order effects
- Platforms with large messaging and social audiences face pressure to offer differentiated AI personas rather than only generic assistants; distribution becomes a key advantage.
- As AI personas appear in feeds and chats, safeguards against manipulation, echo chambers, and harmful dependency become product requirements, not merely model-quality concerns.
Third-order effects
- If users accept synthetic companionship, consumer platforms could increasingly supply conversation and engagement alongside human social connections, changing what a social graph is for.
- That shift would make AI companion relationships a distinct governance category, with scrutiny likely to focus on disclosure, emotional influence, and platform incentives.
The trend: Consumer AI is moving from task-oriented chatbots toward ambient, platform-distributed synthetic companions embedded in everyday social surfaces.