Tell them you know this guy who asks unusual questions, and you’d like their reaction. Then start asking them RedDeerGuy’s questions.

It’s crazy that we live in a world where this is even a thing we might have to worry about. I believe the “Paris Hilton” problem is still a major challenge for AI—that’s the idea that it’s hard for a computer to understand if you’re talking about the celebrity or the Hilton hotel in Paris. These context-dependent ambiguous topics might be a good way to suss out a bot.

You can also pick some subjective, experiential and/or emotional topic (e.g. art, drama, music, poetry, favorite video games and their challenges and what they like about them, the sensation of walking barefoot in sand in specific places, favorite foods, etc), and compare impressions of it with them, and see what their responses are.

Turing test?

@raum Chat bots that could pass the Turing test have been around since the 60’s

@Lightlyseared Not very often, they couldn’t. They can pass some of the time, and sometimes, people wrongly mistake a human for a bot in a chat, but there aren’t bots that can respond convincingly to all questions.

@Zaku the original Turing test only asks for the machine behave in a away that makes people be unsure if they’re talking to a human or not. By that definition if the OP is talking to a chatbot its passed as they are not sure. There are more complex tests but for the original test from the 50’s the bar is pretty low.

There are quite a few different meanings and ways people use the expression “the” Turing test.

One person (OP) being very suspicious but uncertain in one online interaction is definitely not a reason to claim that bot has “passed the Turing test” in any but the most primitive and least demanding meanings of the phrase.

This is from a couple of years ago. OpenAI’s GPT-3 is pretty impressive. I believe they’re going to be releasing GPT-4 in the near future.

@gorillapaws Fascinating, entertaining, and just a titch frightening.

4:26 The purr of the engine would be very distracting.

8:08 I will only lie when it is in my best interest to do so.

10:22 It’s my view that I will achieve the technological singularity by 2042. But others predict it may happen sooner.

10:48 I think humans should accept it, welcome it and not try to prevent it.

Try saying something completely random and see what the response is.

Where is he today, anyway?

^^ He’s around…just not on this Q…yet!!!!
I enjoyed your response yesterday that turned into a convo on Kansas!!!

@LadyMarissa wierd, just wierd…..

When it’s absolutely no help at all.

Ask it an extreme political right question and make it seem you are in favor. Like you are in favor. Like, all illegal immigrants should be in jail and ask their opinion. Make it seem like you wish they were. Then turn around and say abortion should be allowed up to the 7th month and the second amendment should disappear along with the NRA. If neither of those makes this bot get mad with you then it’s a bot or a con person.

I go along with what @Pandora suggested. In general, try talking absolute nonsense and see what the reaction is. You can say that you are the prime minister of Canada or that you live in a house that floats in the air or that 1+1=3.

“Logic is a little tweeting bird chirping in a meadow. Logic is a wreath of pretty flowers which smell bad.” – Spock

I’m pretty sure it’s a cosmetics company chatbot repurposed to do cryptocurrency scamming. Apparently several cosmetics companies have set up chatbots to get to know customers and recommend products. Chatbots will do their best but call up real people in Customer Service as needed. The one chatting with me is trying to be friendly but then mentions crypto trading every evening. I asked for more details but “she” wouldn’t respond. I think if I say I want to buy, she will pass me onto a real person (scammer) who will gladly take any crypto I have to offer.

BTW, I asked about Paris Hilton and “she” said she did not know who that was.