Recent searches
Search options
Stories like this remind me of the Arab Spring. People celebrate the ability of new technology to influence people without asking what happens when it’s used to influence people to think and do things you don’t like.
Experiments that show chatbots can change people’s minds will be looked upon as much of a turning point as Twitter being used to incite protest in the Middle East and GamerGate.
@carnage4life they were the NPCs all along...
@carnage4life Infinite patience is a hell of a power...
Someone was complaining about ChatGPT censorship and I thought "how hard is it to get it to give me Nazi propaganda?" And it was easy... I told it I wanted to write a narrative transcript between Goebbels and a young man, and it would play Goebbels.
I wasn't in danger of being convinced, but I was struck how patient and empathetic this AI Goebbels was. It corrected me very gently, not confronting me or raising my defenses.
@carnage4life
I think many of us could be much better at convincing other people, but we don't put in the work, we don't actively try to explore the other person's perspective, we spend time convincing ourselves or reinforcing our own beliefs instead of engaging with the other person, we won't concede or avoid unimportant side-issues.
An AI can be instructed to do all of those things, and then the damn thing scales. In comparison traditional activists neither train nor scale.
Today in Focus: Black Box: episode 4 – Bing and I https://www.theguardian.com/news/audio/2024/aug/29/black-box-episode-4-bing-and-i-podcast
In the 1st story "when Lee Johnson discovered his wife, Yokie, had cancer, he turned to AI" and received more empathy and care and on-point mental-emotional help in the first minutes, and in hours following, than from months in the medical system.
It's an amazing heart stretching story. One scenario where interaction with generative AI did the right thing, when people were not.
It intimates a possible future where machines provide more care than we do, unless we step up.
@carnage4life This is a reaction to the use of chatbots to influence people to think and do things we don't like, so I don't understand how you think people are unaware of that. That's not to say I think this is a good idea.
@carnage4life @AnarchoNinaWrites That’s really interesting… I wonder, is it that the chatbot comes up with so many factual counterarguments, or because they perceive it as a neutral party? 
@mivox Given that human persuasion is apparently far less effective, it's likely the latter.
That having been said, I retweeted it because I agree with OP; for whatever good this can do, the fact that AI chat bots are capable of persuading people will invariably be used for far more harmful purposes.
@AnarchoNinaWrites Oh absolutely. A very double-edged discovery there. On top of all the other rotten stuff about LLMs. *sigh*
@mivox @AnarchoNinaWrites
Since LLM-driven influencing has the potential to inadvertently harm the target, I'd expect it's more likely to be used by the unscrupulous
@carnage4life If only there were money to be made by stopping the influencing of people altogether and just telling them truth.