Social media platform X (formerly Twitter) recently introduced a new feature called "More Information About This Account," powered by the Grok AI model, but its performance has been disappointing. This feature is reportedly the latest attempt by Elon Musk to test Grok model AI capabilities on the X platform since launching xAI last year.
The feature allows paid users to obtain AI-generated summaries of account information by clicking a button next to the username or display name. However, the actual usage has been less than satisfactory. In multiple tests, Grok either provided overly generic information or completely incorrect data.
For instance, when users inquired about information on colleagues Sean O'Kane and Amanda Silberling, Grok only provided general descriptions of the types of posts they made on the platform. More shockingly, it misidentified another colleague, Jagmeet Singh, as a Canadian expert, despite his infrequent posts on this topic. In another case, Grok even confused an executive from an educational technology company with an Indian cricket player of the same name.
Although X platform included a warning in the feature, alerting users that the Grok 1.5 version might be erroneous, this does not mask the obvious flaws of the feature. Even without providing incorrect information, Grok failed to offer valuable insights beyond the user's public profile.
Considering Grok's access to X platform's data, this new feature should have been able to provide deeper and more accurate user information. However, its current performance falls far short of expectations.
This situation has sparked discussions about the challenges AI faces in practical applications, particularly in terms of information accuracy and utility. Analysts believe that X platform needs to reconsider and improve this feature to fully harness the potential of AI technology and provide truly valuable information for users.