In an attempt to beat competition from Artificial Intelligence (AI) chatbots like OpenAI’s ChatGPT, Google’s Gemini, and Microsoft’s Copilot, Elon Musk has announced his plan to retrain Grok with advanced reasoning.

On his official handle on X (formerly known as Twitter), Musk posted that with the new version he would “rewrite the entire corpus of human knowledge”, remove lacking information, and errors. Users had raised concerns about the reliability of xAI’s chatbot after it showed contradictory information about the Iran-Israel conflict. 

Citing the reason as “Far too much garbage in any foundation model trained on uncorrected data”, he continued, stating that the AI tool would be retrained in this fashion.

Grok’s official page on X carried a post on the new development on June 21, shedding some light on how Musk plans to increase the reliability of the platform: “Determining missing information likely involves Grok 3.5 cross-referencing vast datasets, using logical inference to spot gaps, and reasoning from first principles to hypothesize what should exist.”

The post further stated that, “It may analyze patterns in scientific literature or historical records to identify inconsistencies or sparse areas. Human expert input could validate these gaps, especially in complex fields.” The exact techniques which will be employed are still unclear since Grok 3.5 is still in development.

Grok’s post on X ended with a disclaimer regarding how the scale of rewriting all human knowledge involves risks of bias or errors, and feasibility remains debated.