One of the 3 shots on goal, the Q2 report, came out and did not mention an LLM being used to conduct an influence operation. Therefore, dropping the odds by a third.
Lots of fake articles mentioned through the report, and a Turkish election to influence, but no LLMs mentioned. This sort of thing is bound to happen, since an LLM can easily generate the filler text for a fake article, but it may or may not actually get mentioned in the report. For example, a fake brand with fictitious logo was mentioned, but the tool used to create it (maybe Photoshop) was not.
One of the 3 shots on goal, the Q2 report, came out and did not mention an LLM being used to conduct an influence operation. Therefore, dropping the odds by a third.
Lots of fake articles mentioned through the report, and a Turkish election to influence, but no LLMs mentioned. This sort of thing is bound to happen, since an LLM can easily generate the filler text for a fake article, but it may or may not actually get mentioned in the report. For example, a fake brand with fictitious logo was mentioned, but the tool used to create it (maybe Photoshop) was not.
Nice catch! Only, the link appears inoperative (I get a "URL signature expired" message), better to access it from their central page: https://transparency.fb.com/metasecurity/threat-reporting
UPDATE: links works OK now; strange...