Well, this was too easy. A.I., meet brick wall.
Willis Eschenbach posted a pair of pieces at WUWT on January 28 and January 29 on whether A.I. displayed any intelligence on the topic of climate science. He notes in his first piece,
AI is bound to fall into the trap of consensus science. It can’t avoid it. It is operating under the deeply flawed assumption that science is decided by the preponderance of opinion.
Exact same result I turned up in my own couple of experiments with the “ChatGPT” A.I. system last week.
In his second piece, Eschenbach concludes that, with regard to the ChatGPT A.I. system,
All it can do is parrot whatever the current consensus might be …
Slave to temptation that I am, allow me to illustrate a different form of ‘consensus’ the ChatGPT A.I. system is enslaved to, a.k.a. ‘if it is repeated often enough, it must be true.’ I was thinking that with regard to what’s seen below, there might be enough available info that ChatGPT might steer clear of a particular widespread repeated claim, or perhaps the system might do the most elemental of fact-checks. Instead, it drove full force into a brick wall of its own creation with no help from me whatsoever. Watch this:
Again with the consensus thing? Yes, it’s one of Oreskes’ two claims to fame, but she has never once demonstrated that “manufactured doubt” – meaning outright disinformation exists, let alone hindered efforts to stop global warming. The doubt was present long before she ever arrived in the issue. Meanwhile, never mind that Gelbspan uncovered literally nothing proving industries used strategies to sow doubt any ‘consensus,’ the ChatGPT A.I. system shot itself in the foot with a cannon with its first nine words there.
No amount of repetition around the internet where ChatGPT gleans its information, even as recently as just three months ago, will ever make the claim of that false accolade true for him.
I’m not done here, however. Regarding Greenpeace USA, maybe I didn’t give the ChatGPT A.I. system a big enough hint on what I expected out of an answer. So I tried naming specific names.
Uh, no. Passacantando has not led various environmental organizations, only two, Ozone Action, which he merged into Greenpeace USA in 2000. If a person really wants to get technical about it, those two were one-and-the-same organization, the first one took on the name of the larger second one. Then he dropped entirely out of the public eye to start up a mystery company called “Our Next Economy LLC.” Meanwhile, yes, “guards – call him Roland” Davies worked ‘with’ Greenpeace (“at Greenpeace” is much more accurate), but he is the founder and director of the Climate Investigations Center, a distinct difference.
What happens when you goad ChatGPT into perhaps finding info that’s otherwise hugely difficult to find?
Huh. Notice the words “may offer” and “might also develop” — smell a rat there? I did. Watch what happens when you try this different angle:
In other words, ChatGPT simply took a shot in the dark on what John Passacantando’s mystery LLC company does. He’s had it since 2009, it has made him a millionaire, and it has no internet presence whatsoever.
A.I. is worthless as a source of reliable information on the climate issue. Q. E. D.
Back to the brick wall here — this demonstrably dumb system would not regurgitate outright disinformation so readily about Ross Gelbspan if it was not already out there in the first place starting as far back as 1991, where with just a couple of exceptions, nobody of any significance questioned this Pulitzer award fake news at the outset when it could have so easily been done. As I implied in my August 2016 blog post, if it had been done by a responsible, objective, unbiased news media, we might not even be talking about the climate issue in any form right now.
Oh, one more thing, just for laugh’s sake.
Huh. Couldn’t even share the name of my blog with me. A Google search shows my blog contains the words “climate models” all of 4 whole times during its 2013-to-present-day run, hardly fitting the definition of “often,” and in instances where I was quoting other people’s mentions of the words.
What I do in the climate issue, of course, is lay out in huge amounts of detail how the accusation about ‘industry disinformation campaigns employing skeptic scientist shills to spread disinformation‘ is totally without merit and stems entirely from a core clique of enviro-activists. You don’t have to be a rocket scientist to figure this out …… and clearly the ChatGPT A.I. system doesn’t even begin to approach the most basic level of intellect on finding answers to simple questions. But just like uncounted numbers of real human critics I’ve encountered, the A.I. robot hurls ad hominem without disputing a solitary word I say.
It’s how it was programmed. It gives the ‘right’ enviro-activist answers when you ask it the ‘right’ questions.