Myth: Public Data Is Free for AI to Use

Here's a myth that gets teams in trouble: if it's on the public internet, it's free for AI to use.

Not quite. People share things in a context. A forum post asking for advice. A photo from a family event. A review of a local shop. They shared it with an audience in mind, and "training a commercial AI model" was not that audience.

Scraping it anyway creates real problems. The trust problem: people feel misled when their words show up somewhere they never agreed to. And the compliance problem: privacy frameworks don't give public data a free pass just for being visible.

In plain English: public is not the same as permission.

Before your team trains on scraped data, ask the uncomfortable question: would the people who posted this be okay with this use? If you're not sure, that's your answer.

I explain frameworks; I don't give legal counsel.