Lets have it analyze your question as asked.....

This AI Overview is answering a very different question than the one you've been discussing.

The question it answers is:

"Is it legal to scrape 300,000 photos from a website, to post to your own website, under the guise of fair use?"

If those are the actual facts, then the answer is generally yes, that's a high-risk scenario. Systematically copying an entire image library from another site to create your own competing repository is very unlikely to qualify as fair use.

However, if someone is trying to use this AI Overview to prove that PokerChipper is infringing, it doesn't necessarily fit unless those are actually the facts. The answer assumes several things:

that 300,000 photos were scraped;
that they were copied directly from another website;
that they were republished wholesale;
that the new site is essentially a competing image repository;
and that the operator is relying solely on fair use.

Those assumptions matter.

There are also a few statements in the AI Overview that are too absolute or oversimplified:

It says, "Photos are creative works, which receive the highest level of copyright protection." Photographs are indeed copyrightable works, but courts still analyze photographs under the four fair-use factors. Some landmark fair-use cases involve photographs. There is no rule that photographs are categorically excluded from fair use.
It says, "Copying content just to republish it on a competing website is not transformative." If the use truly is wholesale republication, that's often a strong point against fair use. But courts look at the specific use. There is no blanket rule that any competing use can never be transformative.
It mentions the CFAA (Computer Fraud and Abuse Act). Whether scraping violates the CFAA depends on the facts. After Van Buren v. United States and hiQ Labs v. LinkedIn, the law around scraping publicly accessible websites is more nuanced than the overview suggests. Simply scraping publicly available pages does not automatically violate the CFAA.

One thing the AI Overview doesn't discuss is the public interest in historical documentation. It says nothing about:

historical preservation,
educational reference,
scholarship,
research,
museums,
archives,
documentary uses,
or whether the images are being used as part of a historical catalog rather than simply republished as a gallery.

Those are all considerations that can become relevant depending on the facts.

So I'd caution against treating that AI Overview as an answer to every dispute involving historical reference sites. It's answering a specific hypothetical involving mass scraping and wholesale republication. Whether it applies to any particular website depends on what actually happened.

The broader point remains the same: copyright cases are highly fact-specific. The answer can change significantly depending on the source of the images, the nature of the copying, how the images are used, what rights exist, whether licenses were granted, whether fair use is asserted, and whether the use is educational, historical, documentary, or something else.