We are at an interesting inflection point in how people are using AI to write online.

This week Panagram – the best known AI detection software – released a new free API to detect AI writing on social media feeds.

In practical terms, what that means is that you can easily detect and monitor which posts on Substack, LinkedIn or X are written by AI.

You can expect AI analysis to be incorporated into software the way we have word count in Microsoft Word.

That should lead to a more nuanced view of how and when to use AI.

Instead of being viewed as negative in all cases, it will be acceptable and even preferred for lots of information – like sending directions.

But it will be viewed as inappropriate for communication in which you are trying to make a real connection – everything from a condolence message to seeking funding.

In the short run, expect to be called out for using AI in ways that feel hypocritical or phony.

Like Elizabeth Warren using AI in a post criticizing AI. Or Apple writing its social media post commemorating its 50th anniversary with AI.