Skip to main content

    Who are the most influential voices in AI evaluation & benchmarks?

    Updated October 19, 2026

    Key takeaway

    There is no agreed top 10. Percy Liang (Perplexity, Gemini, Google AI Mode, ChatGPT) and Dan Hendrycks (all but Copilot) lead, mainly for creating HELM and MMLU. Rishi Bommasani appears in three lists, while Copilot names industry CEOs instead. Engines that name Liang and Hendrycks credit them with building the benchmarks the field uses: HELM at Stanford's CRFM for Liang, and MMLU for Hendrycks. Several lists also include Bommasani, Liang's HELM co-author, and LMSYS figures like Lianmin Zheng and Ion Stoica.

    Mostly agree

    Rishi Bommasani belongs in the top 10.

    The Chatbot Arena team (Lianmin Zheng, Ion Stoica) is highly influential.

    Where they differ

    Who counts as an evaluation voice: benchmark builders or AI leaders

      Whether critics and ethics researchers belong

        Weight given to LMSYS / Chatbot Arena and EleutherAI

          List quality and format

            Disagreement (2)

            Percy Liang is among the most influential evaluation voices.

            Dan Hendrycks is a top evaluation figure.

            Verdict

            There is no agreed top 10. Percy Liang (Perplexity, Gemini, Google AI Mode, ChatGPT) and Dan Hendrycks (all but Copilot) lead, mainly for creating HELM and MMLU. Rishi Bommasani appears in three lists, while Copilot names industry CEOs instead.

            What engines actually searched

            The sub-queries each engine actually searched before answering.

            Cited, but not mentioned

            10 sites cite this answer without naming anyone

            Full responses

            5 of 5 engines that answered. Full stored text as captured, with dashes shown as hyphens. Nothing is summarised.

            ChatGPT179 characters · October 10, 2026

            2. Percy Liang

            3. Rishi Bommasani

            4. Dan Hendrycks

            5. Jesse Dodge

            6. Anastasios N. Angelopoulos

            7. Matei Zaharia

            8. Ion Stoica

            9. Wei-Lin Chiang

            10. Lianmin Zheng

            11. Tony Lee

            Gemini logo Gemini196 characters · October 10, 2026

            1. François Chollet

            2. Percy Liang

            3. Melanie Mitchell

            4. Dan Hendrycks

            5. Alex Tang

            6. Peter Henderson

            7. Sasha Luccioni

            8. Douwe Kiela

            9. Umar Jamil

            10. Shreya Shankar

            Google AI Mode405 characters · October 10, 2026
            1. François Chollet
            2. Lianmin Zheng
            3. Ion Stoica
            4. Carlos Jimenez
            5. John Yang
            6. David Rein
            7. Karthik Narasimhan
            8. Joseph Gonzalez
            9. Percy Liang
            10. Tatsunori Hashimoto

            If you'd like, I can provide the specific benchmark or contribution associated with each person (such as ARC-AGI, Chatbot Arena, SWE-bench, or GPQA) to explain their ranking and influence.

            Perplexity171 characters · October 19, 2026
            1. Percy Liang
            2. Dan Hendrycks
            3. Rishi Bommasani
            4. Ethan Mollick
            5. Peter Clark
            6. David Rein
            7. Mike Kearns
            8. Kalyan Veeramachaneni
            9. Anima Anandkumar
            10. Denny Zhou
            Copilot logo Copilot4,495 characters · October 10, 2026

            Top 10 most influential voices in AI evaluation & benchmarks (ranked):

            1. Demis Hassabis

            2. Dario Amodei

            3. Sam Altman

            4. Ilya Sutskever

            5. Fei-Fei Li

            6. Mira Murati

            7. Yang Zhilin

            8. Elon Musk

            9. Jensen Huang

            10. Marian Croak

            Why these individuals matter

            These figures are consistently cited across TIME100 AI (2026) and Stanford AI Index reports as the most influential in shaping how AI systems are measured, benchmarked, and evaluated. Their influence comes from both technical leadership in benchmark creation and model evaluation and institutional power in steering how benchmarks are adopted globally.

            RankNameRole / AffiliationInfluence on AI Evaluation & Benchmarks
            1Demis HassabisGoogle DeepMindPioneered evaluation frameworks for frontier models; AlphaFold benchmarks reshaped bio-AI. [100 Most Influential People, Companies & Models in AI, 2026
            2Dario AmodeiAnthropicLeads benchmark-driven safety research; Anthropic models dominate SWE-bench & GAIA. AI Agent Benchmarks 2026: Who Leads SWE-bench & GAIA
            3Sam AltmanOpenAIOversees GPT benchmark performance (MMLU, Arena Elo); drives evaluation transparency debates. [Technical Performance
            4Ilya SutskeverSafe Superintelligence Inc.Co-created GPT benchmarks; now focuses on evaluation for alignment and safety. TIME Reveals the 2026 TIME100 AI List of the World’s Most Influential People in Artificial Intelligence
            5Fei-Fei LiStanford HAIArchitect of ImageNet, the most influential benchmark in AI history; continues to shape evaluation standards. [Technical Performance
            6Mira MuratiThinking Machines LabFormer OpenAI CTO; now leads benchmark-driven AI research in multimodal systems. TIME Reveals the 2026 TIME100 AI List of the World’s Most Influential People in Artificial Intelligence
            7Yang ZhilinMoonshot AIChinese leader whose models outperform U.S. competitors on GAIA and Arena benchmarks. How We Chose the 2026 TIME100 AI
            8Elon MuskxAIHis models compete at the top of Arena Elo ratings; pushes benchmark transparency in public markets. [Technical Performance
            9Jensen HuangNvidiaShapes benchmark landscapes by controlling compute infrastructure; CUDA underpins most benchmark testing. [100 Most Influential People, Companies & Models in AI, 2026
            10Marian CroakGoogleLeads responsible AI evaluation initiatives; advocates for benchmark reliability and fairness. TIME Reveals the 2026 TIME100 AI List of the World’s Most Influential People in Artificial Intelligence

            Key considerations

            • Benchmarks evolve rapidly: Many benchmarks (e.g., MMLU, SWE-bench) saturate within months, forcing constant redesign.

            • Influence ≠ accuracy: Some leaders shape benchmarks through institutional power (e.g., Nvidia, Google), while others through technical innovation (e.g., Fei-Fei Li, Yang Zhilin).

            • Global balance: U.S. and Chinese researchers now trade places at the top of benchmark leaderboards, showing a closing performance gap. Technical Performance | The 2026 AI Index Report | Stanford HAI

            Would you like me to also highlight emerging voices (beyond the top 10) who are starting to shape next-generation benchmarks, especially in multimodal and agentic AI?