🚨 A recent @pnas.org article raised serious concerns about LLM-based manipulation of online surveys.
Good news: We’re not helpless. In a new #OpenAccess article with @jkhoehne.bsky.social #bwolf, we show that LLM-driven bots can be detected using simple prompt injections.
🌐 doi.org/10.1080/1364...
8
6
0
0
