According to Google Gemini on X, Gemini now surfaces verified scientific citations with direct links to original papers, allowing users to trace claims back to primary sources and strengthen research reliability (source: @GeminiApp, Feb 27, 2026). As reported by the Gemini team, the feature emphasizes high-quality data provenance by linking to publisher and preprint repositories, which can reduce hallucinated references and improve trust in AI-assisted literature reviews (source: @GeminiApp). For businesses, this upgrade enables faster evidence gathering for R&D briefs, regulatory filings, and due diligence workflows by cutting time spent validating sources and enhancing auditability (source: @GeminiApp). According to the announcement, the change positions Gemini for academic search, pharma literature mining, and technical market analysis use cases where verifiable sourcing is critical (source: @GeminiApp).

According to @AcerFur and cited by Greg Brockman (@gdb), a previously unsolved AI research problem now has a newly discovered proof in the literature, which is notably different from earlier methods (source: https://x.com/AcerFur/status/2012770890849689702). KoishiChan located the prior proof, and the result has been updated on the community wiki for transparency. While this does not represent a fully novel result, it highlights the importance of peer review and transparency in AI research. This development underscores the value of revisiting existing literature and community-driven knowledge sharing in accelerating AI theory and algorithm innovation.