WhatsApp has always occupied a different psychological space than other social platforms. No algorithm, no ads, no public performance. Just direct conversations with people you actually know. That intimacy is the product, and this is why Meta, despite owning it since 2014, has been careful not to break what makes it feel different.

Which is what makes this week’s update interesting. The Writing Help feature originally launched last year as a rephrasing and proofreading tool, useful for tightening up a message before you send it, much like having a second pair of eyes. That framing was easy to accept because you still wrote the message, and the AI just helped you say it better.

The new update goes a step further: WhatsApp will now generate suggested replies based on your conversation. You tap an icon, the AI reads the thread, and it offers you something to send back. Meta says the chats remain private even when using this feature.

This is where the product moves from a simple tool to something more complicated. Besides the obvious privacy issues, WhatsApp is positioning this feature as a way to help you get your message “just right,” which implies a more substantive role in what you actually say in high-stake exchanges.

The strategic logic is transparent enough: Meta would strongly prefer that when you reach for AI assistance in a conversation, you use its built-in tool rather than switching to ChatGPT or Gemini and pasting results back in. The company knows people are already doing this, so by building it natively keeps the interaction and the data inside Meta’s ecosystem.

All of which poses the existential question: what happens to a relationship when one side starts outsourcing the thinking?

To learn more, click here.