The ethics violation is definitely bad, but their results are also concerning. They claim their AI accounts were 6 times more likely to persuade people into changing their minds compared to a real life person. AI has become an overpowered tool in the hands of propagandists.
This, of course, doesn’t discount the fact that AI models are often much cheaper to run than the salaries of human beings.
And the fact that you can generate hundreds or thousands of them at the drop of a hat to bury any social media topic in highly convincing ‘people’ so that the average reader is more than likely going to read the opinion that you’re pushing and not the opinion of the human beings.
The ethics violation is definitely bad, but their results are also concerning. They claim their AI accounts were 6 times more likely to persuade people into changing their minds compared to a real life person. AI has become an overpowered tool in the hands of propagandists.
It would be naive to think this isn’t already in widespread use.
I mean that’s the point of research: to demonstrate real world problems and put it in more concrete terms so we can respond more effectively
deleted by creator
And the fact that you can generate hundreds or thousands of them at the drop of a hat to bury any social media topic in highly convincing ‘people’ so that the average reader is more than likely going to read the opinion that you’re pushing and not the opinion of the human beings.