AI belongs in the research. Not in the byline.

Not for being AI-generated. Google’s stated position is that it rewards helpful content however it was produced, while its spam policies target content generated at scale primarily to manipulate rankings. In practice that means unedited bulk AI publishing is at risk and carefully edited AI-assisted work is not.

I use it for research, clustering, analysis and first drafts. Everything published is rewritten by me, fact-checked against sources, and carries my name as reviewer. I tell clients this openly, because the alternative — charging for human writing and delivering generated text — is simply dishonest.

Yes, and plenty does. The failure mode is not detection, it is quality: generated content tends to be generic, occasionally wrong, and free of the specifics and opinions that make a page worth citing. It ranks until something better exists, which on any commercially interesting topic is soon.

Check the facts. Generated content states plausible things confidently, so errors cluster around numbers, dates and specifics. Also look for content that is fluent but says nothing you could not have guessed — no prices, no process, no opinion, no constraints.

It has made research and analysis much cheaper and made genuinely good content marginally cheaper at best, because the expensive part was always judgement and verification. Anyone offering dramatically cheaper content because of AI is passing on the saving from the step they stopped doing.

Google does not require it. I think naming a human author and reviewer is the more useful signal, because it says somebody is accountable for the content being right — which is the actual question a reader or an engine is asking.