Skip to content

Everything that could go wrong with X's new AI written community notes

    If AI Note writers “generate initial concepts that represent a broader range of perspectives than a single human writer who could usually be improved, the quality of the community's consultation has improved from the start,” the newspaper said.

    Future of AI Notes

    Researchers imagine that as soon as the testing of X was completed, AI Note writers could not only help in investigating problematic messages marked by human users, but also one day selected messages that are expected to become viral and cannot spread more information than human reviewers could.

    Additional benefits of this automated system, they suggested, would include X NOTE rating agents who have quick access to more thorough research and evidence synthesis, as well as a clearer note of the notes that can accelerate the assessment process.

    And perhaps in a day AI agents could even learn to predict the assessment scores to accelerate things even more, researchers speculated. However, more research would be needed to ensure that this would not homogenize communitynotities, so that they are so well polished that nobody reads them.

    Perhaps the most muskian ideas of ideas presented in the newspaper is an idea of ​​training AI Note writers with clashing views to “debate opponent about the merits of a note.” Probably that “could help immediately potential errors, hidden prejudices or manufactured evidence, so that the human assessor can make a better informed judgment.”

    “Instead of starting all over again, the rater now plays the role of a referee – evaluating a structured collision of arguments,” the newspaper said.

    Although X may be moving to reduce the workload for X users who write Community Notes, it is clear that AI could never replace people, researchers said. Those people are needed for more than just ai-written notes stamped with rubber.

    Human tones that “have been written all over again” are valuable to train the AI ​​agents and the niche expertise of some rater cannot be easily replicated, the newspaper said. And perhaps most clear, people “are uniquely positioned to identify shortages or prejudices” and are therefore forced to write notes “on topics that overlook the automated writers,” such as spam or scams.