This question actually involves three different authorities, and they give three different answers, so the clearest way through is to take them one at a time.

Google's answer is no. Search has no labeling requirement, and adding a label does not hurt your rankings, because the algorithms evaluate the quality of a page rather than who or what wrote it. Ahrefs' large-scale studies confirm this indifference in practice, since 86.5 percent of top-twenty pages already contain some AI content and their million-page follow-up concluded that Google punishes bad content rather than AI content. What Google does police is thin content produced at scale, and that is a quality standard rather than a disclosure standard.

The law's answer changed six weeks ago, which is why this question is trending. Article 50 of the EU AI Act has applied since August 2, 2026, with penalties reaching 15 million euros or 3 percent of worldwide turnover, and it covers non-EU businesses when their content is used in the Union. California's SB 942 is moving in the same direction.

The detail that matters most for a blog is the exemption built into the text rules. The Act exempts content that has undergone human review or editorial control, where a natural or legal person holds editorial responsibility for the publication, and the review has to be genuine rather than a rubber stamp. A store whose AI-drafted posts pass through a real editor who approves them and stands behind them sits inside that exemption, which is one more argument for a draft-mode workflow where a human clears every post before it goes live.

Fully automated, unreviewed AI text aimed at EU readers is a different situation, and there disclosure does apply. The same goes for anything that could be mistaken for human when it is not, which is why social formats carry the highest practical risk.

Synthetic media is treated more strictly than text. AI-generated or manipulated images, audio and video need a perceptible disclosure, where a small visible badge is enough but buried fine print or an invisible watermark alone is not. The EU has published optional official icons for exactly this purpose, and the major image generators now embed C2PA provenance metadata that you should preserve through your publishing pipeline.

The third authority is your reader, and that answer depends on context. Nobody expects a disclosure on a product description, but on content that trades on personal experience, an undisclosed AI byline becomes a trust problem if it is discovered, while a plain sentence about how you work usually reads as confidence. Whatever you decide about labels, the content itself still has to clear the helpfulness bar, because a disclosure neither rescues thin content nor penalizes good content.

Pulled together, the answer looks like this: Google never requires a label, EU-facing synthetic media and unreviewed AI text have required one since August 2026, and reviewed blog posts with real editorial ownership fall under the exemption. Since the review workflow is something you should run anyway for quality, your compliance posture and your content quality end up being the same habit. When the stakes are real, ask a lawyer rather than a blog post.