Or someone could train a bot to exaggerate misinformation from known liars
Community note: In fact, Oregon sent over 420 fire trucks but due to the rampage of the woke chupacabra and the cannibals of the northern California mountains, only 69 made it as far south as Sacramento. At that point California officials realized that Oregon actually sent minivans full of weed and spent the weekend hotboxing, using the Jewish space laser that started the fires on the “defrost” setting to heat up the marijuana.
dont need AI for this. its why we have systems like webs of trust and peer verification. standard crypto + some social communication solves a lot of these problems.
first question you’d need to answer is how joe-bob got into the web of trust in the first place. in order for them to abuse their position they needed to be added to your web.
literally what bluesky is doing with their filters.
if you want to automate it… run statistical analysis on the user for how many people ban them. you eventually find the people who are heavy but accurate with the ban button and can use them as inputs. and can use them as early gauges for new accounts. combine that with account age and you’ll eventually get a fairly robust and automated platform for banning misinformation accounts.
This is something AI would be good for
Have it search for specific misinformation and reply with a canned response with sources using an official account.
Or someone can just train an AI to spit out more misinformation.
And so, the bot wars started
We’re years in at this point.
Then it’s going to be AI vs AI
In this case the misinformation AI will be run by Musk’s sugar daddy Putin so it won’t get banned
Or someone could train a bot to exaggerate misinformation from known liars
Community note: In fact, Oregon sent over 420 fire trucks but due to the rampage of the woke chupacabra and the cannibals of the northern California mountains, only 69 made it as far south as Sacramento. At that point California officials realized that Oregon actually sent minivans full of weed and spent the weekend hotboxing, using the Jewish space laser that started the fires on the “defrost” setting to heat up the marijuana.
dont need AI for this. its why we have systems like webs of trust and peer verification. standard crypto + some social communication solves a lot of these problems.
How can a web of trust prevent Joe-bob from claiming bullshit like in the OP?
first question you’d need to answer is how joe-bob got into the web of trust in the first place. in order for them to abuse their position they needed to be added to your web.
literally what bluesky is doing with their filters.
if you want to automate it… run statistical analysis on the user for how many people ban them. you eventually find the people who are heavy but accurate with the ban button and can use them as inputs. and can use them as early gauges for new accounts. combine that with account age and you’ll eventually get a fairly robust and automated platform for banning misinformation accounts.
Have you seen some of the “factual” information AI states definitively?
Generic LLMs are shit, they try to do everything