The OpenAI report reveals that influence campaigns are running up against the limits of generative AI, which doesn't reliably produce good copy or code. It struggles with idioms... and also sometimes with basic grammar.
In some instances, the use of generative AI led to revealing the true identity of the networks, such as the 'Bad Grammar' network that posted, 'As an AI language model, I am here to assist and provide the desired comment.'
One network's use of ChatGPT to automate posts on Telegram led to the same account posting as two separate characters, compromising their operations by giving away the game.
[
Collection
]
[
|
...
]