Social platform x will pilot feature This allows AI chatbots to returns the community records.
Community records are a Twitter feature that has been resolved by the elon of the service, now called X. US users who can signify a particular post, which is then checked by another user before appearing to be mounted in posts. The community note may appear, for example, in video post that is not clear about synthetic origins, or as addendum to the wrong positism.
The note is public when you get the consensus between a group that disagrees in the past.
The community note is quite successful on x to inspiration Meta, Tiktokand Youtube To pursue the same initiative – meta Abolish The third check-party program is exchange for lower costs, and public labor.
But it remains visible that the use of Chatbots AI is the reality-check can prove or dangerous.
These AI records can be produced using grible x or using other AI tools and connect to x via API. Any note that AI should be considered the same as the notes sent by the person, which means that it will pass the same vetting process to support the accuracy.
The use of AI in reality-check is unfillable, given in general to AIS to AIS Hallucinateor make the context not based on the fact.
According to the paper Published This week by researchers who are working on the Community Note X, suggested humanity and llms can be used in tandem. The human feedback can add a note of note through the reinforcement, with a sense of human note is still the last check before the note is published.
“The purpose is not to make a AI assistant that tells the user, but to build a ecosystem that strengthen the human to think of critical and knowing the world better,” Paper said. “Llms and humans can work together on a good loop.”
Even by checking humanity, still risk to trust in AI, especially since the user will be able to contain third party llms. ChattPT opening, for example, a problem that has just experienced a model SycophanticSee rankings-. If LLM Prioorities “Help” through completing the facts, then the comments made ai can stop being unacceptable.
There is also concerned that Rater will be too much from the comments made from AI, lowering the motivation to complete this volunteer work.
Users are not expecting the community note that made – X plans to test these AI contributions for a few weeks before you play back if you succeed.
