Much ado these days with people thinking “A.I.” is some kind of all-encompassing savior to make life easier. Just do a generic Google search of “A.I. can help with” and watch the system fill in the last words with a whole range of situations … with Google A.I.’s own automatically generated input right at the very top, where its handy helpful little links in its “Show More” section to expand what it offers. All as though “A.I.” is benign, soulless and without political bias.

Speaking of Google searches, I’ve been using its basic system almost exclusively for over a decade, since that system clearly head & shoulders above any other search engine – you just have to know how to circumvent its biased results by doing boolean searches to prompt results it might not want the public to know about. It’s how I discovered exactly what the Ozone Action environmental group was, and who its staff were. But in May 2024, Google began adding its ‘AI option’ to its search methods menu. I’ve avoided it like the plague, knowing and proving already just months earlier that at least some forms of A.I. had no intelligence whatsoever. For this blog post, however, I’ll actually look into Google’s “A.I. Overview” for the very first time.  Watch this —

“Who discovered the fossil fuel industry memos with the phrase Reposition Global Warming as Theory (Not Fact)? What is their importance?”

Regarding the screencaptures below, let me first say that when I’ve shown any of Google search screencaptures previously, my images included the url address so that readers could take those addresses and see for themselves what the pages look like in full context. But I always deleted its auto-generated url section revealing what my web browser is – that’s nobody’s business. In the following exercise, I did so in a second tab (for the purpose of making the sceencaptures an optimal size) . . . and was amazed to see Google hugely change just the first paragraph in the blink of an eye. Special emphasis here on the red-highlighted words:

First iteration paragraph, with the url containing my browser ID:

The “Reposition Global Warming as Theory (Not Fact)” memo was revealed in 1991 when investigative journalists obtained documents detailing a campaign strategy by the Information Council on the Environment (ICE), a fossil-fuel industry front group …

Versus the second iteration, generated less than a minute afterward:

The memos with the phrase “Reposition Global Warming as Theory (Not Fact)” were not discovered by a single individual but were publicly revealed in 1998 by the oil industry watchdog group Ozone Action. The specific memo containing this phrase was from the American Petroleum Institute (API), outlining a communications plan for climate deception.

My God.  In the first paragraph iteration, at least Google’s artificial “I” got the date right, but no investigative journalists ever obtained the memo, it was leaked by enviro-activists to any receptive news outlet they could find. And as usual, an indicator throughout the years that anyone hurling accusations about the memo and related docs were regurgitating talking points fed to them without ever actually seeing the docs, the official name of the ICE campaign was “Information Council FOR the Environment.

But it’s that second seconds-later paragraph iteration that’s the clincher.

• the “1998” revelation date is totally wrong.
• Ozone Action was solely originally devoted to saving the ozone layer, only later in 1995 did they switch to the more lucrative issue of global warming. They never self-identified in any manner as an “oil industry watchdog group.”
• the “reposition global warming” memo set had exactly zip, zero, nada association with the American Petroleum Institute communications plan – a.k.a., the notorious “victory will be achieved” memo. Regarding that last monster error — Hold . That . Thought . until down near the end of this blog post.

This fault with Google’s auto-generated “A.I. Overview” doesn’t stop there. Try dropping my pair of gotcha questions into a Google search window for yourselves and see what kind of different result you get. For laugh’s sake, I tried a third attempt. I kid you not, the thing spit out this:

• Kert Davies created his Climate Investigations Center in 2014, after he left Greenpeace. But as seen in his LinkedIn resumé, he did work at Ozone Action back in the late 1990s, which did somehow magically ‘obtain’ the “reposition global warming memos and other docs, but they never said from whom.
• The Desmogblog spit out their accusation about the memo – from what I saw at the time – way back in 2013. Albeit still in johnny-come-lately fashion, since Al Gore ‘published’ the phrase full screen in his movie in 2006 and the New York Times published the phrase in 1991.

My question posed to Google about the importance of the “Reposition global warming” memo set is a gotcha question – no matter which “A.I. Overview” result you get, it attempts to tell the public that the phrase is ‘smoking gun’ proof of industry disinformation campaigns. As I’ve already demonstrated myriad ways here at GelbspanFiles, the memo set had no importance whatsoever; the thing was never implemented in any form anywhere (obtuse suggestions for audience targeting included). The group it was sent to hadn’t even solicited it, and their copy ended up in the trash. Don’t expect any version of A.I. “assistance” to tell you that.

I tried one more attempt with “A.I. Overview”, to see what would happen:

Hilarious.  The people who’d tell you Ross Gelbspan / Ross Gelbspan / Ross Gelbspan unearthed / uncovered / discovered the memo would say Google’s “A.I. Overview” can be trusted exactly as far as you can throw the whole system. I’d tell you those same individuals and others like them can be trusted on their own assertions equally far. Gelbspan – as I’ve already detailed – was also a johnny-come-lately on promulgating the accusation about the memo no earlier than November 1995.

All of this – I should note – without ever diving straight into Google’s pure “A.I. Mode” (I tried it briefly for this exercise here. The info result was different enough from the examples above, that it would be a topic for a whole separate GelbspanFiles blog post).

What their “A.I. Overview” provides is not help to the overall public, it’s only helpful to the core clique of enviro-activists who need to keep the false accusation alive about the memos being proof that the fossil fuel industry hired ‘shill’ scientists to spread disinformation to undercut the ‘settled science’ of fossil fuel usage causing catastrophic global warming.

How is it proven that A.I. has no actual intelligence, that it cannot actually think? Ask it questions and/or give it a task to accomplish, where you know what the correct answers are. UK Journalist/book author Steve Boggan neatly illustrated how this is done: “I asked an AI assistant to write my Wikipedia entry — it couldn’t stop lying.”

This was also entertainingly illustrated concerning one of my own writings just recently when WUWT reproduced [  ahem  reproduced! ] my ‘deep into the details weeds’ 8/1/25 GelbspanFiles blog post where – to their disadvantage – commenters were reluctant to dive into the details, (one to the point of ironic unfairness), even when prompted by the value of it by one of the other commenters. One particular person there decided A.I. could “help” summarize what he was unwilling to absorb for himself.

See what happened there? His “Grok A.I.” did exactly what the ‘Google A.I. second iteration’ result did that I showed up at near the top of this post; it put the “reposition global warming” memo set into direct association with the totally unrelated “victory will be achieved” memo. As I showed in a comment reply to him, the “Grok A.I.” summary he got provided him with information that was dead wrong — I know, because I’ve had the correct information for years now. The “victory will be achieved” memo is the second-most effective accusation element among the four totally false elements seen in the “ExxonKnew” lawsuits.

Bottom line: Rely on “A.I.” to provide you with information, and the roadmap it gives you to follow may direct you right off a cliff.
————————————————————————
One more thing: Anybody hazard a guess as to why one of my blog posts is an item at an A.I. “machine learning” website?