I worry a little bit that people who refuse to learn about ai as a part of their anti ai position are going to be extremely unprepared to understand what’s actually scary about it and already have their digital literacy at risk tbh
not that I am some genius in this regard but if you follow ai developments even slightly you might change the things you are most worried about. do people know the extent to which ai is already eating itself and how meaningless this is making swaths of the internet. do people know that there are plenty of random mid-sized companies today buying their employees’ likenesses to create digital clones and using these to make hundreds of videos. I am so much more worried about labor and surveillance and abuse than people becoming lazy about writing emails. and idk man I sort of like and respect people who are willfully ignorant about it as a way of minimizing its force in their lives and I am in some ways jealous but also when I see posts that basically still boil down to “chatgpt will never fool me” I am like 😭😭😭 for one thing not the only thing to be concerned about, for another thing I am really sorry but I don’t think you’re right
Exactly. I think about this a lot when I see "ChatGPT is always confidently wrong" arguments, not because genAI is often right, but because the argument doesn't work on people who actually use it regularly.
It isn't always wrong. It's often correct enough for the average person's day to day usage. If you say, "it always gives misinformation", they're going to ignore you because just that morning it gave a perfectly functional summary of their meeting's notes. It correctly identified some vegan restaurants in their area. It accurately calculated some math for them.
If it was always wrong, no one would use it. The danger is that you can never tell when it's going to be wrong, and when it's right enough that people start to trust it and not double check anything (automation bias), the things it gets wrong become catastrophic. Sometimes automation bias is so bad that mistakes compound and proliferate for ages before anyone catches it (ie: the company who were making financial decisions based off of hallucinated forecasting).
See, my particular bugaboo about AI is the overwhelming mountain of evidence that it LITERALLY CAUSES BRAIN DAMAGE:
MIT Study Finds Artificial Intelligence Use Reprograms the Brain, Leading to Cognitive Decline
Cognitive Atrophy Paradox of AI–Human Interaction: From Cognitive Growth and Atrophy to Balance
The Silent Erosion: Global Generational Cognitive Decline in the Age of AI and the Future of Human Intellectual Agency
How AI is rewiring the human brain: the generational transformation of cognition and knowing
AI-overdependence and human cognitive decline: Hazards, evidence, and mitigation strategies
Is AI dulling our minds? — Harvard Gazette
Your Brain on AI: Cognitive Offloading, Debt, and Atrophy
Is Generative AI Rewiring Our Brains? Here's How It Happens
I think this is an easier argument to make with the people in your life who use AI. "It's wrong enough of the time that you shouldn't trust it" is easy to refute with "but it's right just enough of the time to be worth it". Add to that the good old Dunning-Kruger effect, and it's easy to give AI way too much credit. And way too much power.
Instead of asking them to vet AI's accuracy (especially about things they don't understand themselves), ask them about what effect AI has had on their own brain. Have they noticed changes in their memory, their attention span, their ability to focus?
Ask them what they use AI for. Ask them how they did those things before they had AI - and now ask them if they could do that now.
Ask them to consider what it would be like if they didn't use AI for a month, or a week, or even a day.
If the thought makes them uncomfortable, beg them to consider why.
I work in memory care, so believe me, I know what the process of dementia often looks like, and I'm seeing the telltale signs in more and more people who are younger and younger than ever. The scariest part is that the average diagnosis usually comes several years after the first symptoms present. It worries me.
And, as others in this thread have mentioned, there are even signs of cognitive decline in the AI itself as llm models inevitably begin to eat their own tails:
Almost all leading AI chatbots show signs of cognitive decline
Older AI models show signs of cognitive decline, study shows
Digital Dementia? AI Shows Surprising Signs of Cognitive Decline
Oh, so we're all getting less and less functional. Cool. Cool.
