Google just updated its spam policies to specifically target "recommendation poisoning" — attempts to game AI answers by stuffing content with signals designed to manipulate AI citations.

The announcement was quiet. The implication is not.

For the past two years, a small industry has grown up around Generative Engine Optimization: tactics designed to make your content appear in AI-generated answers by whatever means necessary. Some of it has been legitimate — genuine authority building, original research, structured data. Some of it has been the usual search-optimization shortcut culture applied to a new platform.

Google's message suggests the shortcut phase is ending faster than expected.

This pattern is entirely familiar. Every significant platform in the history of digital marketing has followed the same arc: open and manipulable → gamed aggressively → locked down with prejudice against the gamers. Google's core search went through it. Facebook reach went through it. Email deliverability went through it.

AI search optimization is following the same arc, just faster.

The companies that built genuine authority through original research, distinctive expertise, and earned media coverage will find themselves in better shape as the algorithm tightens. The companies that built AI visibility through synthetic signals may discover those signals have become a liability.

The lesson available in every previous version of this story: the techniques that survive platform maturation are the ones that would still make sense even if the algorithm changed tomorrow.

Building real authority always passes that test. Gaming the algorithm rarely does.