[suthener]

“This story is a good example that in some cases AI can be quite stupid.”

I asked Grok recently to list my state’s representatives and party. It listed one as being in office who had lost his primary. That’s really pathetic. Now I don’t trust it for anything.

[redfreedom]

IMO AI is no different than a human student. Students learn what’s put in front of them and what they are given access to.

I often ask Google’s AI questions. When asked a political question, it gives politically slanted answers. This is to be expected from Google, which affirms my thoughts above.

And there are fools that want to put some of our national defense assets into the hands of AI. They need to watch the movie War Games.

[bgill]

Don’t use that one. Been seeing a lot of problems with it.

[Political Junkie Too]

They can have their glitches sometimes for unexplainable reasons.

I was researching the claims of some articles recently and I asked a rather lengthy question of the Perplexity Pro AI, which is currently my favorite. In the question, I referenced several recent events in the news for the AI to consider in its response:

1. The Trump-Zelensky blow-up in the White House.
2. The Mamdani primary win in New York.

In both cases, the AI said that there was no credible information that either of those events actuall happened!

I challenged the AI that they actually did happen and it then confirmed that they were indeed true events. It couldn't give me a good explanation as to why it missed those facts the first time around.

-PJ