{
    "componentChunkName": "component---src-gatsby-pdf-topic-pdf-tsx",
    "path": "/pdf/topics/artificial-intelligence",
    "result": {"data":{"platform":{"topic":{"id":"65c55d599e947c438698b7a7","slug":"artificial-intelligence","name":"Artificial Intelligence","path":"/topics/artificial-intelligence","__typename":"Platform_Topic","created":"2021-07-26T04:37:47.00","published":null,"doiId":null,"outlineNumber":"1.1","trend":{"id":"65c55d5c9e947c438698b86d","path":"/trends/quantum-revolution-advanced-ai","slug":"quantum-revolution-and-advanced-ai","name":"Quantum Revolution & Advanced AI","__typename":"Platform_Trend","outlineNumber":"1","openGraph":{"image":{"thumbnails":{"card":{"url":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_480/v1760333329/gesda-platform/banner/banner-trend-1_image__P1A_2026_nqu5xa.webp"}}}},"color":{"id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"}},"description":{"text":"Recent progress stems from machine learning (ML), which enables computers to discover patterns in data.[2](/citation/2025-01-1-1-2/) Within ML, deep learning — neural networks loosely inspired by the brain — has driven most breakthroughs.[3](/citation/2025-01-1-1-3/) These models now match or surpass human performance in many tasks: image recognition (2015), winning at Go (2016), predicting protein structures (2022) and enabling chatbots with powerful language skills.\n\nThe rise of generative AI has further transformed the field. Unlike earlier systems, which primarily classified or predicted, generative models create new text, images, music, code and even video. Tools such as chatbots, image generators and coding assistants are now integrated into search engines and office software, with early “agents” capable of performing economically useful tasks with limited supervision.\n\nThis progress has been enabled by abundant data, vast computing power and model designs that scale. But challenges remain: these systems are opaque, energy-intensive and prone to bias, misinformation and “hallucinations.” Generative AI also raises concerns about intellectual property, misinformation and impacts on education, creative industries and the future of work.\n\nNew approaches are needed for AI that is efficient, adaptable and trustworthy. Meanwhile, powerful AI is already here, and society must prepare for its far-reaching consequences in how humans create, reason and interact with machines.\n\n**KEY TAKEAWAYS**\n\nThe field of AI has seen rapid progress in recent years, with particularly significant advances being seen in AI that generates text, images and video based on learning from online data. However, there is great scope for further advances. Large language models (LLMs), for example, embody cultural knowledge but are inadequate in causal reasoning and robustness. Building on current success, the **Future of generative AI** appears to be in ever-wider applications and dealing with architectural gaps: transformer architectures are powerful for predicting tokens and pixels, but inadequate for processing continuous, high-dimensional natural signals in the real world. Progress in AI also requires grounding models in physical reality beyond text: **World-modelling and embodied AI** are vital for true intelligence. This will enable an AI system to incorporate representations of a physical environment, complete with physical parameters that might constrain or otherwise influence its output. Intelligence emerges from embodiment and active interaction with the environment, drawing from biological and developmental insights. Greater consideration of real-world situations is likely to help **AI for science**, which is currently growing in its influence on the scientific endeavour. AI shows great promise for scientific discovery but faces challenges in scientific data generation and abstraction. Future advances depend on more efficient data use and new learning paradigms that move beyond brute-force scaling. Breakthroughs in **AI foundations** may be necessary to fulfil the potential currently seen in the field, while AI safety and alignment with human values are paramount for societal trust and beneficial deployment: considering the broader social and educational implications of AI is critical for responsible progress."},"intro":{"text":"Artificial Intelligence (AI) aims to produce intelligence using algorithms and machines. This includes systems that perceive and analyse their environment, take decisions, communicate and learn. In the past decade, AI has reached major milestones and is poised to disrupt societal norms.[1](/citation/2025-01-1-1-1/)"},"anticipatoryImpact":{"text":"Three fundamental questions guide GESDA’s mission and drive its work: Who are we, as humans? How can we all live together? How can we ensure the well-being of humankind and the sustainable future of our planet? We asked researchers from the field to anticipate what impact future breakthroughs could have on each of these dimensions. This wheel summarises their opinions when considering each of these questions, with a higher score indicating high anticipated impact, and vice versa.\n\n* Anticipated impact on who we are as humans\n* Anticipated impact on how we will all live together\n* Anticipated impact on the well-being of humankind and sustainable future of our planet"},"indicatorValues":[{"id":"65c55cf49e947c438698aa55","value":"0.541","numericValue":0.541,"year":2024,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}},{"id":"68a45e5ba44e85ae4b8f05d3","value":"0.561","numericValue":0.561,"year":2025,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}},{"id":"6a9fa2b3670cd190fd135717","value":"4.5","numericValue":4.5,"year":0,"indicator":{"id":"6a9f993c670cd190fd1356df","name":"Impact on People","title":"Impact on People","slug":"impact-on-people","dataSetId":null,"color":{"id":"6a9fa21e670cd190fd1356f8","value":"#fcbe0d"}}},{"id":"6a9fa2d7670cd190fd13571f","value":"3.5","numericValue":3.5,"year":0,"indicator":{"id":"6a9f994c670cd190fd1356e3","name":"Impact on Planet","title":"Impact on Planet","slug":"impact-on-planet","dataSetId":null,"color":{"id":"6a9fa245670cd190fd135700","value":"#83a599"}}},{"id":"6a9fa2e8670cd190fd135724","value":"6.5","numericValue":6.5,"year":0,"indicator":{"id":"6a9f995c670cd190fd1356e7","name":"Impact on Society","title":"Impact on Society","slug":"impact-on-society","dataSetId":null,"color":{"id":"6a9fa236670cd190fd1356fc","value":"#d88c73"}}}],"editions":[{"id":"66ab1bb636a8f2f336a557bf","name":"2024","slug":"2024","numericValue":2024},{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025","numericValue":2025}],"anticipatoryImpactImage":{"image":{"id":"image_gesda-platform/image-asset/psp-pl-1-25-1-1_image__PSP-PL1_25_1.1_dpms8i","url":"https://res.cloudinary.com/shapeable/image/upload/v1760069218/gesda-platform/image-asset/psp-pl-1-25-1-1_image__PSP-PL1_25_1.1_dpms8i.webp"}},"embeds":{"citations":[{"id":"691a7999c0043bba84a9e5c3","slug":"2025-01-1-1-1","url":"https://openreview.net/pdf?id=BZ5a1r-kVsf","name":"OpenReview.net","authors":[{"id":"691a77e2c0043bba84a9d9f3","name":"Y. LeCun","slug":"y-le-cun"}],"authorShowsEtAl":null,"edition":null,"publication":"OpenReview.net","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":1,"year":null},{"id":"691a79f3c0043bba84a9e833","slug":"2025-01-1-1-2","url":"https://www.springerprofessional.de/en/machine-learning-methods/26496716","name":"Machine Learning Methods","authors":[{"id":"66f4d48609a10d3d0e1494ea","name":"H. Li","slug":"h-li"}],"authorShowsEtAl":null,"edition":null,"publication":"Springer","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":2,"year":null},{"id":"691a79aec0043bba84a9e65b","slug":"2025-01-1-1-3","url":"https://doi.org/10.1038/nature14539","name":"Deep learning","authors":[{"id":"691a77e4c0043bba84a9d9fb","name":"Y. LeCun et al.","slug":"y-le-cun-et-al"}],"authorShowsEtAl":null,"edition":null,"publication":"Nature","accessDate":null,"startPage":436,"volume":521,"footnoteNumber":3,"year":null}],"imageAssets":[]},"surveyObservations":{"text":"Within the sub-topics covered by Artificial Intelligence, the **Future of generative AI** is perceived to be already mature, and there is a high awareness about its effects. Even though the field is seen as transformative, its high maturity and relatively clear impact limits its Anticipation Potential scores. Significant advances in **World-modelling and embodied AI**, on the other hand, are still close to 10 years away and highly uncertain. Finally, work on **AI foundations** is likely to continue to evolve and impact other fields. **AI for science** is the topic where the need for multilateral cooperation is highest in order to drive progress and harness future opportunities. "},"color":{"id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993","darkValue":"#4a2851","veryDarkValue":"#2a172e"},"banner":{"id":"6a922a266860f5861ece0731","name":"\"A Network Takes Shape\" by Yamamoto Taiyo, University of Zurich","description":{"text":"\"A Network Takes Shape\" by Yamamoto Taiyo, University of Zurich"},"image":{"id":"image_gesda-platform/banner/a-network-takes-shape-by-yamamoto-taiyo-university-of-zurich_image__22A_Network_Takes_Shape_22_by_Yamamoto_Taiyo_University_of_Zurich_rtfh7h","url":"https://res.cloudinary.com/shapeable/image/upload/v1787963911/gesda-platform/banner/a-network-takes-shape-by-yamamoto-taiyo-university-of-zurich_image__22A_Network_Takes_Shape_22_by_Yamamoto_Taiyo_University_of_Zurich_rtfh7h.jpg","thumbnails":{"mainBanner":{"url":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_1440/v1787963911/gesda-platform/banner/a-network-takes-shape-by-yamamoto-taiyo-university-of-zurich_image__22A_Network_Takes_Shape_22_by_Yamamoto_Taiyo_University_of_Zurich_rtfh7h.jpg","url2x":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_2880/v1787963911/gesda-platform/banner/a-network-takes-shape-by-yamamoto-taiyo-university-of-zurich_image__22A_Network_Takes_Shape_22_by_Yamamoto_Taiyo_University_of_Zurich_rtfh7h.jpg"}}}},"chartImage":{"id":"65c55cee9e947c438698a95a","slug":"chart-1-1-advanced-artificial-intelligence","image":{"id":"image_gesda-22/image-asset/chart-1-1-advanced-artificial-intelligence_image__TRR-1_1-TBC-01","url":"https://res.cloudinary.com/shapeable/image/upload/v1668986362/gesda-22/image-asset/chart-1-1-advanced-artificial-intelligence_image__TRR-1_1-TBC-01.png","url2x":null}},"citations":[{"__typename":"Platform_Citation","_schema":{"label":"Citation","pluralLabel":"Citations"},"typeLabel":"Report","slug":"2025-01-1-1-1","url":"https://openreview.net/pdf?id=BZ5a1r-kVsf","name":"OpenReview.net","authors":[{"id":"691a77e2c0043bba84a9d9f3","name":"Y. LeCun","slug":"y-le-cun"}],"authorShowsEtAl":null,"edition":null,"publication":"OpenReview.net","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":1,"year":null},{"__typename":"Platform_Citation","_schema":{"label":"Citation","pluralLabel":"Citations"},"typeLabel":"Website","slug":"2025-01-1-1-2","url":"https://www.springerprofessional.de/en/machine-learning-methods/26496716","name":"Machine Learning Methods","authors":[{"id":"66f4d48609a10d3d0e1494ea","name":"H. Li","slug":"h-li"}],"authorShowsEtAl":null,"edition":null,"publication":"Springer","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":2,"year":null},{"__typename":"Platform_Citation","_schema":{"label":"Citation","pluralLabel":"Citations"},"typeLabel":"Website","slug":"2025-01-1-1-3","url":"https://doi.org/10.1038/nature14539","name":"Deep learning","authors":[{"id":"691a77e4c0043bba84a9d9fb","name":"Y. LeCun et al.","slug":"y-le-cun-et-al"}],"authorShowsEtAl":null,"edition":null,"publication":"Nature","accessDate":null,"startPage":436,"volume":521,"footnoteNumber":3,"year":null}],"subTopics":[{"id":"65c55d4f9e947c438698b665","name":"Future of generative AI","path":"/sub-topics/future-of-generative-ai","outlineNumber":"1.1.1","slug":"future-of-generative-ai","__typename":"Platform_SubTopic","color":{"id":"65c55cbc9e947c438698a322","name":"Purple","value":"#966993"},"topic":{"id":"65c55d599e947c438698b7a7","slug":"artificial-intelligence","path":"/topics/artificial-intelligence"},"intro":{"text":"The most widely appreciated progress in AI has come from models able to generate new content, such as text, images, video, audio and software code. Generative AI, as the field is known, is underpinned by extremely large models trained on enormous text and image datasets scraped from the internet."},"description":{"text":"Current generative AI research is heavily reliant on brute-force scaling with ever-larger datasets and computing power, raising questions about sustainability and future data limitations. There is a call for architectural innovation enabling AI to learn efficiently from modest data,[4](/citation/2025-01-1-1-4/) drawing from concepts in human learning and cognitive science. \n\nApproaches to cope with data limitations include generating synthetic data, employing new data diversification algorithms[5](/citation/2025-01-1-1-5/),[6](/citation/2025-01-1-1-6/) and developing smarter, data-efficient learning paradigms. However, progress in generative AI must overcome issues with robustness and generalisation by moving from “vanilla” data synthesis to creating data that does not yet exist online, and by exploring alternative architectures to transformers. The field must also avoid over-fixation on current trends (brute-force scaling) and foster unconventional collaborations, especially between academia and industry, to drive fundamental change. \n\nIt is important to emphasise that ethical considerations, safety and alignment of generative models with societal values should be foundational to further adoption.[7](/citation/2025-01-1-1-7/) It is also important to keep technical progress aligned with societal needs and to raise critical questions about long-term future directions of the field.[8](/citation/2025-01-1-1-8/)"},"anticipationScores":{"text":"The Anticipation Potential of a research field is determined by the capacity for impactful action in the present, considering possible future transformative breakthroughs in a field over a 25-year outlook. A field with a high Anticipation Potential, therefore, combines the potential range of future transformative possibilities engendered by a research area with a wide field of opportunities for action in the present. We asked researchers in the field to anticipate: \n\n1. The *uncertainty* related to future science breakthroughs in the field\n2. The *transformative* *effect* anticipated breakthroughs may have on research and society\n3. The *scope for action* in the present in relation to anticipated breakthroughs. \n\nThis chart represents a summary of their responses to each of these elements, which when combined, provide the *Anticipation Potential* for the topic. See [methodology](/science-anticipation/methodology) for more information."},"anticipationScoresImage":{"id":"68e3746fcf743e6f2486699c","image":{"id":"image_gesda-platform/image-asset/1-1-1-sub-anti-2026_image__1.1.1_sub_anti_2026_f5bdfz","url":"https://res.cloudinary.com/shapeable/image/upload/v1759736874/gesda-platform/image-asset/1-1-1-sub-anti-2026_image__1.1.1_sub_anti_2026_f5bdfz.webp","url2x":null,"width":1200,"height":1200}},"horizons":[{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a73d","name":"1.1.1 - 5-year horizon","slug":"1-1-1-5-year-horizon","intro":{"text":"Generative models continue to improve"},"description":{"text":"Research achieves continued improvement of generative models using current architectures (such as transformers), focusing on better data efficiency and robustness. There is widespread use of synthetic data generation to supplement limited datasets. Significant advances in alternative learning paradigms beyond brute-force scaling are made, with new architectures appearing that enable learning from smaller, more diverse datasets. AI models begin to synthesise truly novel data, expanding creativity and generalisation."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89a","name":"5-year horizon","slug":"5-year-horizon","years":5,"title":"5-year","subtitle":"horizon"},"embeds":{"citations":[]}},{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a73e","name":"1.1.1 - 10-year horizon","slug":"1-1-1-10-year-horizon","intro":{"text":"Generative AI incorporates different kinds of information"},"description":{"text":"Generative AI incorporates multimodal and grounded information more effectively, potentially integrating embodied learning and world models. Ethical alignment mechanisms reach practical deployment, fostering greater societal trust."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89b","name":"10-year horizon","slug":"10-year-horizon","years":10,"title":"10-year","subtitle":"horizon"},"embeds":{"citations":[]}},{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a73f","name":"1.1.1 - 25-year horizon","slug":"1-1-1-25-year-horizon","intro":{"text":"Generative-AI systems display adaptive creativity"},"description":{"text":"Further progress enables generative-AI systems to display adaptive creativity and self-supervised scenario generation, facilitating breakthroughs in science, the arts and society. Fundamental innovations reshape collaborative research paradigms between academia and industry."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89c","name":"25-year horizon","slug":"25-year-horizon","years":25,"title":"25-year","subtitle":"horizon"},"embeds":{"citations":[]}}],"indicatorValues":[{"id":"65c55cf49e947c438698aa78","value":"0.496","numericValue":0.496,"year":2024,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}},{"id":"68edcb3faf9e6d6d63270b7d","value":"0.520","numericValue":0.52,"year":2025,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}}],"embeds":{"citations":[{"slug":"2025-01-1-1-4","url":"https://arxiv.org/abs/2508.12680","name":"Towards General Vision Language Reasoning with Multi-Domain Data Curation","authors":[{"name":"Y. Zha et al."}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":4,"year":null},{"slug":"2025-01-1-1-5","url":"https://arxiv.org/abs/2305.06500","name":"InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning","authors":[{"name":"W. Dai et al."}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":5,"year":null},{"slug":"2025-01-1-1-6","url":"https://arxiv.org/abs/2203.06386","name":"Enabling Multimodal Generation on CLIP via Vision-Language Knowledge Distillation","authors":[{"name":"W. Dai et al."}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":6,"year":null},{"slug":"2025-01-1-1-7","url":"https://www.scientificamerican.com/article/how-to-build-an-empathetic-robot/","name":"How To Build An Empathetic Robot","authors":[{"name":"P. Fung"}],"authorShowsEtAl":null,"edition":null,"publication":"Scientific American","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":7,"year":null},{"slug":"2025-01-1-1-8","url":"https://arxiv.org/abs/2501.17805","name":"International AI Safety Report","authors":[{"name":"Y. Bengio et al."}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":8,"year":null}],"imageAssets":[]}},{"id":"65c55d4f9e947c438698b688","name":"World-modelling and embodied AI","path":"/sub-topics/world-modelling-and-embodied-ai","outlineNumber":"1.1.2","slug":"world-modelling-and-embodied-ai","__typename":"Platform_SubTopic","color":{"id":"65c55cbc9e947c438698a322","name":"Purple","value":"#966993"},"topic":{"id":"65c55d599e947c438698b7a7","slug":"artificial-intelligence","path":"/topics/artificial-intelligence"},"intro":{"text":"Despite their impressive capabilities, leading AI models understand relatively little about the world around them. This is because the text and image data they are trained on contains only indirect knowledge of physical reality. This limits their ability to both understand and act in dynamic real-world environments.\n"},"description":{"text":"Internal world models, which allow an agent to predict environmental states given actions, are central to intelligence, supporting zero-shot generalisation and planning.[9](/citation/2025-01-1-1-9/) Embodied intelligence, which is intelligence that is rooted in the connection between perception, action and reality, can be likened to evolutionary and child-development processes: human children’s learning, viewed as “scientists in the crib”, provides a model for active exploration and self-directed experiment in intelligent systems. However, current text-based models lack the necessary grounding and are orders of magnitude less data-efficient than humans. Real-world, multimodal data is critical for building future models, and robotics offers a promising path: here, the integration of language, vision, memory and manipulation capabilities supports hierarchical planning and lifelong learning.[10](/citation/2025-01-1-1-10/)\n\nIn the long term, AI robots may develop complex self-awareness and seamlessly acquire new skills from rich sensory experience, moving closer to human-like capabilities. Safety boundaries must be set, though. For example,  reinforcement-\tlearning strategies that incentivise pure survival should be avoided to ensure that AI acts as a tool aligned with human interests. Some researchers feel that highly abstract reasoning and planning must always be reducible to linguistic or symbolic representations in order to maintain transparency, but there is debate about this. There is also debate about whether LLM-based approaches can generalise to richly structured, non-textual reality such as images and 3D worlds. Overall, embodied learning will be important for credible AI advancement."},"anticipationScores":{"text":"The Anticipation Potential of a research field is determined by the capacity for impactful action in the present, considering possible future transformative breakthroughs in a field over a 25-year outlook. A field with a high Anticipation Potential, therefore, combines the potential range of future transformative possibilities engendered by a research area with a wide field of opportunities for action in the present. We asked researchers in the field to anticipate: \n\n1. The *uncertainty* related to future science breakthroughs in the field\n2. The *transformative* *effect* anticipated breakthroughs may have on research and society\n3. The *scope for action* in the present in relation to anticipated breakthroughs. \n\nThis chart represents a summary of their responses to each of these elements, which when combined, provide the *Anticipation Potential* for the topic. See [methodology](/science-anticipation/methodology) for more information."},"anticipationScoresImage":{"id":"68e8906e63d1c853e9788b94","image":{"id":"image_gesda-platform/image-asset/1-1-2-sub-anti-2026_image__1.1.2_sub_anti_2026_xaybe1","url":"https://res.cloudinary.com/shapeable/image/upload/v1760071745/gesda-platform/image-asset/1-1-2-sub-anti-2026_image__1.1.2_sub_anti_2026_xaybe1.webp","url2x":null,"width":1200,"height":1200}},"horizons":[{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a740","name":"1.1.2 - 5-year horizon","slug":"1-1-2-5-year-horizon","intro":{"text":"Basic multimodal data is embodied in simulation environments"},"description":{"text":"Researchers integrate basic multimodal data (vision, sound and language) with rudimentary world models in embodied simulation environments. Early advances in robotic reasoning and planning with limited grounding in reality."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89a","name":"5-year horizon","slug":"5-year-horizon","years":5,"title":"5-year","subtitle":"horizon"},"embeds":{"citations":[]}},{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a741","name":"1.1.2 - 10-year horizon","slug":"1-1-2-10-year-horizon","intro":{"text":"AI agents acquire more powerful world models"},"description":{"text":"AI agents acquire deeper hierarchical world models, capable of richer zero-shot generalisation and more advanced planning in open-ended tasks. Robotics platforms provide lifelong learning through continual sensory experience."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89b","name":"10-year horizon","slug":"10-year-horizon","years":10,"title":"10-year","subtitle":"horizon"},"embeds":{"citations":[]}},{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a742","name":"1.1.2 - 25-year horizon","slug":"1-1-2-25-year-horizon","intro":{"text":"Embodied AI achieves human-like capabilities"},"description":{"text":"Embodied AI achieves human-like capabilities in manipulation and exploration, including advanced self-awareness and real-time learning. Grounding in physical reality and multimodal perception becomes standard, facilitating robust causal reasoning. Autonomous AI systems operate across diverse domains, dynamically adapting and self-experimenting to discover new skills and strategies. World models move towards being as flexible and abstract as those seen in biological intelligence, driving new forms of interaction between intelligent systems and the environment."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89c","name":"25-year horizon","slug":"25-year-horizon","years":25,"title":"25-year","subtitle":"horizon"},"embeds":{"citations":[]}}],"indicatorValues":[{"id":"65c55cf49e947c438698aab6","value":"0.539","numericValue":0.539,"year":2024,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}},{"id":"68edd0f7af9e6d6d63270bc7","value":"0.580","numericValue":0.58,"year":2025,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}}],"embeds":{"citations":[{"slug":"2025-01-1-1-9","url":"https://arxiv.org/abs/2507.05169","name":"Critiques of World Models","authors":[{"name":"E.Xing et al."}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":9,"year":null},{"slug":"2025-01-1-1-10","url":"https://arxiv.org/abs/2506.22355","name":"Embodied AI Agents: Modeling the World","authors":[{"name":"P. Fung et al."}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":10,"year":null}],"imageAssets":[]}},{"id":"65c55d4f9e947c438698b64d","name":"AI for science","path":"/sub-topics/ai-for-science","outlineNumber":"1.1.3","slug":"ai-for-science","__typename":"Platform_SubTopic","color":{"id":"65c55cbc9e947c438698a322","name":"Purple","value":"#966993"},"topic":{"id":"65c55d599e947c438698b7a7","slug":"artificial-intelligence","path":"/topics/artificial-intelligence"},"intro":{"text":"One of the most exciting possibilities of AI lies in its potential to accelerate scientific discovery. But how and where to apply AI to science requires careful consideration.[11](/citation/2025-01-1-1-11/)"},"description":{"text":"All science involves abstracting data into representations that make future data predictable. AI research, especially in world modelling, can be helpful to science if it focuses on learning these abstractions rather than reconstructing every detail. AI, as a “digital microscope”, provides novel, data-driven approaches to scientific problems, differing from classical rationalism and reductionism. Scientific discovery is currently constrained by the pace and cost of data generation, and AI can accelerate progress via automated experiment design and high-throughput simulation.[12](/citation/2025-01-1-1-12/) AI will also democratise tools for scientific discovery, allowing for new digital experiments, hypothesis generation and global-scale simulations such as digital organisms and weather models.\n\nAI may itself become a target of scientific study in terms of new mathematical frameworks and methods such as “vector-based probability”. There are open challenges still: one is achieving integration across abstraction levels — from language to vision, for instance, and from virtual to physical domains. Nevertheless, there is optimism that AI-driven approaches will create new scientific paradigms and domains, though modesty is warranted, as fundamental principles of intelligence and discovery remain undiscovered.\n"},"anticipationScores":{"text":"The Anticipation Potential of a research field is determined by the capacity for impactful action in the present, considering possible future transformative breakthroughs in a field over a 25-year outlook. A field with a high Anticipation Potential, therefore, combines the potential range of future transformative possibilities engendered by a research area with a wide field of opportunities for action in the present. We asked researchers in the field to anticipate: \n\n1. The *uncertainty* related to future science breakthroughs in the field\n2. The *transformative* *effect* anticipated breakthroughs may have on research and society\n3. The *scope for action* in the present in relation to anticipated breakthroughs. \n\nThis chart represents a summary of their responses to each of these elements, which when combined, provide the *Anticipation Potential* for the topic. See [methodology](/science-anticipation/methodology) for more information."},"anticipationScoresImage":{"id":"68e8a1e363d1c853e9788ccc","image":{"id":"image_gesda-platform/image-asset/1-1-3-sub-anti-2026_image__1.1.3_sub_anti_2026_ai9qdi","url":"https://res.cloudinary.com/shapeable/image/upload/v1760076243/gesda-platform/image-asset/1-1-3-sub-anti-2026_image__1.1.3_sub_anti_2026_ai9qdi.webp","url2x":null,"width":1200,"height":1200}},"horizons":[{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a743","name":"1.1.3 - 5-year horizon","slug":"1-1-3-5-year-horizon","intro":{"text":"AI accelerates scientific discovery in data-rich fields"},"description":{"text":"AI accelerates scientific discovery in data-rich fields such as astronomy and genomics. Such fields experience accelerated discovery thanks to AI-driven automated analysis and experiment design. More nuanced integrative models linking text, images and simulations are developed."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89a","name":"5-year horizon","slug":"5-year-horizon","years":5,"title":"5-year","subtitle":"horizon"},"embeds":{"citations":[]}},{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a744","name":"1.1.3 - 10-year horizon","slug":"1-1-3-10-year-horizon","intro":{"text":"AI-driven science becomes routine in some areas"},"description":{"text":"High-throughput research and complex simulations routinely make use of AI to facilitate digital experimentation and hypothesis generation. New methodologies are underpinned by novel mathematical representation."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89b","name":"10-year horizon","slug":"10-year-horizon","years":10,"title":"10-year","subtitle":"horizon"},"embeds":{"citations":[]}},{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a745","name":"1.1.3- 25-year horizon","slug":"1-1-3-25-year-horizon","intro":{"text":"AI-curated scientific disciplines emerge"},"description":{"text":"As AI systems autonomously generate and test theories, simulate virtual organisms and link abstract reasoning to physical experimentation, AI becomes the curator of numerous scientific disciplines. Interdisciplinary AI collaboration creates new paradigms and research domains. Scientific progress reaches unprecedented levels, with AI becomes both a central research tool and object of scientific inquiry itself. AI systems begin to autonomously contribute foundational principles to scientific knowledge.\n"},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89c","name":"25-year horizon","slug":"25-year-horizon","years":25,"title":"25-year","subtitle":"horizon"},"embeds":{"citations":[]}}],"indicatorValues":[{"id":"65c55cf49e947c438698aa54","value":"0.550","numericValue":0.55,"year":2024,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}},{"id":"68edd130af9e6d6d63270bd6","value":"0.560","numericValue":0.56,"year":2025,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}}],"embeds":{"citations":[{"slug":"2025-01-1-1-11","url":"https://polymathic-ai.org","name":"Advancing Science through MultiDisciplinary AI","authors":[{"name":"Polymathic"}],"authorShowsEtAl":null,"edition":null,"publication":"Polymathic-AI.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":11,"year":null},{"slug":"2025-01-1-1-12","url":"https://arxiv.org/abs/2408.08873","name":"Accelerating Giant Impact Simulations with Machine Learning","authors":[{"name":"C. Lammers et al."}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":12,"year":null}],"imageAssets":[]}},{"id":"65c55d4f9e947c438698b64e","name":"AI foundations","path":"/sub-topics/ai-foundations","outlineNumber":"1.1.4","slug":"ai-foundations","__typename":"Platform_SubTopic","color":{"id":"65c55cbc9e947c438698a322","name":"Purple","value":"#966993"},"topic":{"id":"65c55d599e947c438698b7a7","slug":"artificial-intelligence","path":"/topics/artificial-intelligence"},"intro":{"text":"Deep learning has dramatically accelerated progress on a wide range of AI problems. This has led to large swathes of the research community focusing their efforts on refining and scaling existing approaches to tackle increasingly complex tasks. But the underlying technology remains in its infancy and continued progress may require fundamental conceptual breakthroughs."},"description":{"text":"LLMs and correlated AI models, for example, excel at fluency and generalisation, but are prone to spurious correlations, lack causal and counterfactual reasoning, and can be fooled by perceptual illusions. Robust intelligence will require new representations and architectures, going beyond scaling and focusing on joint perception-action models, internal consistency and intervention-centric learning.[13](/citation/2025-01-1-1-13/),[14](/citation/2025-01-1-1-14/),[15](/citation/2025-01-1-1-15/) If research is to progress in helpful ways, there is a need to rethink education,[16](/citation/2025-01-1-1-16/) both in the content of statistics and algorithms, and in interdisciplinary skills spanning philosophy, ethics and computational sciences.[17](/citation/2025-01-1-1-17/) Ethical, safety and alignment considerations must also be embedded at the foundational level, especially regarding long-term impacts on society, democracy and human flourishing.\n\nThe future of AI is as much about discovering new principles as it is about continuous improvement and scaling of current architectures. Researchers in this field must remain humble and open to unexpected discoveries. Furthermore, AI’s trajectory should be guided not only by intellectual elegance but also by societal needs, human values and democracy."},"anticipationScores":{"text":"The Anticipation Potential of a research field is determined by the capacity for impactful action in the present, considering possible future transformative breakthroughs in a field over a 25-year outlook. A field with a high Anticipation Potential, therefore, combines the potential range of future transformative possibilities engendered by a research area with a wide field of opportunities for action in the present. We asked researchers in the field to anticipate: \n\n1. The *uncertainty* related to future science breakthroughs in the field\n2. The *transformative* *effect* anticipated breakthroughs may have on research and society\n3. The *scope for action* in the present in relation to anticipated breakthroughs. \n\nThis chart represents a summary of their responses to each of these elements, which when combined, provide the *Anticipation Potential* for the topic. See [methodology](/science-anticipation/methodology) for more information."},"anticipationScoresImage":{"id":"68e8a02f63d1c853e9788cab","image":{"id":"image_gesda-platform/image-asset/1-1-4-sub-anti-2026_image__1.1.4_sub_anti_2026_jzravy","url":"https://res.cloudinary.com/shapeable/image/upload/v1760075811/gesda-platform/image-asset/1-1-4-sub-anti-2026_image__1.1.4_sub_anti_2026_jzravy.webp","url2x":null,"width":1200,"height":1200}},"horizons":[{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a748","name":"1.1.4 - 25-year horizon","slug":"1-1-4-25-year-horizon","intro":{"text":"AI technology and societal interfaces are reshaped"},"description":{"text":"AI models embedded with alignment to human values and democratic ideals become standard. Philosophy and ethics become core elements in AI research and deployment. The field moves towards integrating cognitive, ethical and computational sciences. AI is ever more tightly integrated into learning, governance and societal advancement."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89c","name":"25-year horizon","slug":"25-year-horizon","years":25,"title":"25-year","subtitle":"horizon"},"embeds":{"citations":[]}},{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a747","name":"1.1.4 - 10-year horizon","slug":"1-1-4-10-year-horizon","intro":{"text":"Foundational principles undergo fundamental shifts"},"description":{"text":"New representations and abstraction layers replace correlation-centric models, altering the way AI systems are developed and the ways in which they can be deployed. AI systems demonstrate deeper causal reasoning and counterfactual inference, with education systems adapting to teach new forms of algorithmic literacy."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89b","name":"10-year horizon","slug":"10-year-horizon","years":10,"title":"10-year","subtitle":"horizon"},"embeds":{"citations":[]}},{"__typename":"Platform_Horizon","id":"65c55ce69e947c438698a746","name":"1.1.4 - 5-year horizon","slug":"1-1-4-5-year-horizon","intro":{"text":"Advances in AI architectures address robustness and causal reasoning"},"description":{"text":"Increased focus on intervention-centric and perception-action models yield more robust AI that displays causal reasoning. Early integration of ethical and safety constraints at the architectural level creates an ethos of responsible development."},"color":{"__typename":"Platform_Color","id":"65c55cbc9e947c438698a322","name":"Purple","slug":"purple","value":"#966993"},"type":{"__typename":"Platform_HorizonType","id":"65c55ce79e947c438698a89a","name":"5-year horizon","slug":"5-year-horizon","years":5,"title":"5-year","subtitle":"horizon"},"embeds":{"citations":[]}}],"indicatorValues":[{"id":"65c55cf49e947c438698aa59","value":"0.577","numericValue":0.577,"year":2024,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}},{"id":"68edd185af9e6d6d63270be5","value":"0.590","numericValue":0.59,"year":2025,"indicator":{"id":"65c55cf29e947c438698aa4d","name":"Anticipation Potential","title":null,"slug":"anticipation-potential","dataSetId":"ANTICIPATION_POTENTIAL","color":null}}],"embeds":{"citations":[{"slug":"2025-01-1-1-13","url":"https://arxiv.org/abs/2110.07178","name":"Symbolic Knowledge Distillation: from General Language Models to Commonsense Models","authors":[{"name":"P. West et al."}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":13,"year":null},{"slug":"2025-01-1-1-14","url":"https://arxiv.org/abs/2505.13111","name":"Why Knowledge Distillation Works in Generative Models: A Minimal Working Explanation","authors":[{"name":"S. Cha and K. Cho"}],"authorShowsEtAl":null,"edition":null,"publication":"ArXiv.org","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":14,"year":null},{"slug":"2025-01-1-1-15","url":"Https://doi.org/10.1109/JPROC.2021.3058954","name":"Toward Causal Representation Learning","authors":[{"name":"B. Schölkopf et al."}],"authorShowsEtAl":null,"edition":null,"publication":"Proceedings of the IEEE","accessDate":null,"startPage":612,"volume":109,"footnoteNumber":15,"year":null},{"slug":"2025-01-1-1-16","url":"https://educationaldatamining.org/edm2025/proceedings/2025.EDM.workshop-tutorial-abstracts.250/index.html","name":"Bringing Generative Artificial Intelligence (GenAI) to Education","authors":[{"name":"H. Li et al."}],"authorShowsEtAl":null,"edition":null,"publication":"Educational Data Mining Conference 2025","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":16,"year":null},{"slug":"2025-01-1-1-17","url":"https://mitpress.mit.edu/9780262037310/elements-of-causal-inference","name":"Elements of Causal Inference Foundations and Learning Algorithms","authors":[{"name":"J. Peters et al."}],"authorShowsEtAl":null,"edition":null,"publication":"MIT Press","accessDate":null,"startPage":null,"volume":null,"footnoteNumber":17,"year":null}],"imageAssets":[]}}],"committeeMembers":[{"__typename":"Platform_Person","id":"65c55d249e947c438698b086","name":"Alexandre Pouget","slug":"alexandre-pouget","position":"Professor","path":"/people/alexandre-pouget","badge":"2024","color":{"value":"#966993"},"topicExpertise":[{"id":"66cbf7cf0177877dc528d4a9","name":"||Anticipation Committee Member|Advanced AI","slug":"anticipation-committee-member-advanced-ai-4","editions":[{"id":"66ab1bac36a8f2f336a557ba","name":"2023","slug":"2023"},{"id":"66ab1bb636a8f2f336a557bf","name":"2024","slug":"2024"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b0a1","name":"Antoine Bosselut","slug":"antoine-bosselut","position":"Global Expert","path":"/people/antoine-bosselut","badge":"2022","color":{"value":"#966993"},"topicExpertise":[{"id":"65c55d5b9e947c438698b7cf","name":"||Anticipation Committee Member|Advanced AI","slug":"anticipation-committee-member-advanced-ai-6","editions":[{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"6849511563371e51d83bdef3","name":"Bernhard Schölkopf","slug":"bernhard-scholkopf","position":"Director, Department of Empirical Inference; Scientific Director, ELLIS Institute Tübingen","path":"/people/bernhard-scholkopf","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"6849511263371e51d83bdeec","name":"Bernhard Schölkopf||Anticipation Committee Member|Artificial Intelligence","slug":"bernhard-scholkopf-anticipation-committee-member-artificial-intelligence","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b0b7","name":"Bin Yu","slug":"bin-yu","position":"Global Expert","path":"/people/bin-yu","badge":"2022","color":{"value":"#966993"},"topicExpertise":[{"id":"65c55d5b9e947c438698b7d8","name":"Bin Yu||Anticipation Committee Member|Advanced AI","slug":"bin-yu-anticipation-committee-member-advanced-ai","editions":[{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b0b8","name":"Blaise Agüera y Arcas","slug":"blaise-aguera-y-arcas","position":"Vice-President and Google Fellow","path":"/people/blaise-aguera-y-arcas","badge":"2024","color":{"value":"#966993"},"topicExpertise":[{"id":"66c7de88f99a0ea32cd1f84a","name":"||Anticipation Committee Member|Advanced AI","slug":"anticipation-committee-member-advanced-ai","editions":[{"id":"66ab1bac36a8f2f336a557ba","name":"2023","slug":"2023"},{"id":"66ab1bb636a8f2f336a557bf","name":"2024","slug":"2024"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b0fc","name":"Emmanuel Abbé","slug":"emmanuel-abbe","position":"Chair of Mathematical Data Science / Full Professor","path":"/people/emmanuel-abbe","badge":"2024","color":{"value":"#966993"},"topicExpertise":[{"id":"65c55d5b9e947c438698b7eb","name":"||Anticipation Committee Chair|Advanced AI","slug":"anticipation-committee-chair-advanced-ai-1","editions":[{"id":"66ab1bac36a8f2f336a557ba","name":"2023","slug":"2023"},{"id":"66ab1bb636a8f2f336a557bf","name":"2024","slug":"2024"},{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"}],"topic":{"id":"65c55d599e947c438698b7a7"}},{"id":"66cbf7420177877dc528d491","name":"||Anticipation Committee Member|Advanced AI","slug":"anticipation-committee-member-advanced-ai-3","editions":[{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"684951b663371e51d83bdf2a","name":"Eric Xing","slug":"eric-xing","position":"Founding President","path":"/people/eric-xing","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"684951c963371e51d83bdf34","name":"Eric Xing||Anticipation Committee Member|Advanced AI","slug":"eric-xing-anticipation-committee-member-advanced-ai","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"68494d9563371e51d83bdeae","name":"Hang Li","slug":"hang-li","position":"Head of Research","path":"/people/hang-li","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"68494d9163371e51d83bdea7","name":"Hang Li||Anticipation Committee Member|Artificial Intelligence","slug":"hang-li-anticipation-committee-member-artificial-intelligence","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b15c","name":"Jennifer Chayes","slug":"jennifer-chayes","position":"Global Expert","path":"/people/jennifer-chayes","badge":"2022","color":{"value":"#966993"},"topicExpertise":[{"id":"65c55d5b9e947c438698b80d","name":"Jennifer Chayes||Anticipation Committee Member|Advanced AI","slug":"jennifer-chayes-anticipation-committee-member-advanced-ai","editions":[{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"6849509063371e51d83bdecb","name":"Jitendra Malik","slug":"jitendra-malike","position":"Arthur J. Chick Professor of Electrical Engineering and Computer Sciences","path":"/people/jitendra-malike","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"6849508d63371e51d83bdec4","name":"Jitendra Malik||Anticipation Committee Member|Artificial Intelligence","slug":"jitendra-malik-anticipation-committee-member-artificial-intelligence","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b16b","name":"John C. Platt","slug":"john-c-platt","position":"Global Expert","path":"/people/john-c-platt","badge":"2022","color":{"value":"#966993"},"topicExpertise":[{"id":"65c55d5b9e947c438698b814","name":"||Anticipation Committee Member|Advanced AI","slug":"anticipation-committee-member-advanced-ai-7","editions":[{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"684ac2f72b92aded195bba7b","name":"Jun Zhu","slug":"jun-zhu","position":"Bosch AI Professor, Computer Science Department","path":"/people/jun-zhu","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"684ac2f32b92aded195bba74","name":"Jun Zhu||Anticipation Committee Member|Artificial Intelligence","slug":"jun-zhu-anticipation-committee-member-artificial-intelligence","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"68494ae863371e51d83bddfa","name":"Kyunghyun Cho","slug":"kyunghyun-cho","position":"Professor of Computer Science and Data Science; Co-Director of Global Frontier AI Lab,","path":"/people/kyunghyun-cho","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"68494c3463371e51d83bde3b","name":"Kyunghyun Cho||Anticipation Committee Member|Advanced AI","slug":"kyunghyun-cho-anticipation-committee-member-advanced-ai","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"6849516963371e51d83bdf10","name":"Maosong Sun","slug":"maosong-sun","position":"Professor Department of Computer Science and Technology","path":"/people/maosong-sun","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"6849516663371e51d83bdf09","name":"Maosong Sun||Anticipation Committee Member|Artificial Intelligence","slug":"maosong-sun-anticipation-committee-member-artificial-intelligence","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b231","name":"Samy Bengio","slug":"samy-bengio","position":"Global Expert","path":"/people/samy-bengio","badge":"2024","color":{"value":"#966993"},"topicExpertise":[{"id":"65c55d5b9e947c438698b851","name":"||Anticipation Committee Member|Advanced AI","slug":"anticipation-committee-member-advanced-ai-5","editions":[{"id":"66ab1bac36a8f2f336a557ba","name":"2023","slug":"2023"},{"id":"66ab1bb636a8f2f336a557bf","name":"2024","slug":"2024"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"68494c0263371e51d83bde30","name":"Shirley Ho","slug":"shirley-ho","position":"Research Professor of Physics","path":"/people/shirley-ho","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"68494bfc63371e51d83bde29","name":"Shirley Ho||Anticipation Committee Member|Artificial Intelligence","slug":"shirley-ho-anticipation-committee-member-artificial-intelligence","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b23c","name":"Sébastien Bubeck","slug":"sebastien-bubeck","position":"Senior Principal Research Manager","path":"/people/sebastien-bubeck","badge":"2024","color":{"value":"#966993"},"topicExpertise":[{"id":"66c7df55f99a0ea32cd1f863","name":"||Anticipation Committee Member|Advanced AI","slug":"anticipation-committee-member-advanced-ai-1","editions":[{"id":"66ab1bac36a8f2f336a557ba","name":"2023","slug":"2023"},{"id":"66ab1bb636a8f2f336a557bf","name":"2024","slug":"2024"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"68494d2063371e51d83bde8f","name":"Yann LeCun","slug":"yann-le-cun","position":"VP & Chief AI Scientist","path":"/people/yann-le-cun","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"68494d1d63371e51d83bde88","name":"Yann LeCun||Anticipation Committee Member|Artificial Intelligence","slug":"yann-le-cun-anticipation-committee-member-artificial-intelligence","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"68494b6263371e51d83bde12","name":"Yejin Choi","slug":"yejin-choi","position":"Dieter Schwarz Foundation HAI Professor, Professor of Computer Science, and Senior Fellow","path":"/people/yejin-choi","badge":"2025","color":{"value":"#966993"},"topicExpertise":[{"id":"68494c5563371e51d83bde4a","name":"Yejin Choi||Anticipation Committee Member|Advanced AI","slug":"yejin-choi-anticipation-committee-member-advanced-ai","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]}],"committeeChairs":[{"__typename":"Platform_Person","id":"65c55d249e947c438698b0fc","name":"Emmanuel Abbé","slug":"emmanuel-abbe","position":"Chair of Mathematical Data Science / Full Professor","path":"/people/emmanuel-abbe","openGraph":{"id":"openGraph_person/emmanuel-abbe","title":"Emmanuel Abbé","image":{"id":"image_gesda-22/person/emmanuel-abbe_photo__t2_c_yq1mfs","url":"https://res.cloudinary.com/shapeable/image/upload/v1668987088/gesda-22/person/emmanuel-abbe_photo__t2_c_yq1mfs.jpg","url2x":null,"thumbnails":{"card":{"id":"thumbnails-card-file_gesda-22/person/emmanuel-abbe_photo__t2_c_yq1mfs","url":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_480/v1668987088/gesda-22/person/emmanuel-abbe_photo__t2_c_yq1mfs.jpg","url2x":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_960/v1668987088/gesda-22/person/emmanuel-abbe_photo__t2_c_yq1mfs.jpg"}}}},"organisation":{"id":"65c55d059e947c438698ac3b","name":"EPFL","slug":"epfl"},"topicExpertise":[{"id":"65c55d5b9e947c438698b7eb","name":"||Anticipation Committee Chair|Advanced AI","slug":"anticipation-committee-chair-advanced-ai-1","editions":[{"id":"66ab1bac36a8f2f336a557ba","name":"2023","slug":"2023"},{"id":"66ab1bb636a8f2f336a557bf","name":"2024","slug":"2024"},{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"}],"topic":{"id":"65c55d599e947c438698b7a7"}},{"id":"66cbf7420177877dc528d491","name":"||Anticipation Committee Member|Advanced AI","slug":"anticipation-committee-member-advanced-ai-3","editions":[{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"66c3d8782e2f8caf2cc79ebd","name":"Pascale Fung","slug":"pascale-fung","position":"Senior Director of AI Research","path":"/people/pascale-fung","openGraph":{"id":"openGraph_person/pascale-fung","title":"Pascale Fung","image":{"id":"image_gesda-platform/person/pascale-fung_photo__Pascale_Fung_tvua8v","url":"https://res.cloudinary.com/shapeable/image/upload/v1724112439/gesda-platform/person/pascale-fung_photo__Pascale_Fung_tvua8v.jpg","url2x":null,"thumbnails":{"card":{"id":"thumbnails-card-file_gesda-platform/person/pascale-fung_photo__Pascale_Fung_tvua8v","url":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_480/v1724112439/gesda-platform/person/pascale-fung_photo__Pascale_Fung_tvua8v.jpg","url2x":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_960/v1724112439/gesda-platform/person/pascale-fung_photo__Pascale_Fung_tvua8v.jpg"}}}},"organisation":{"id":"68494c6e63371e51d83bde57","name":"Meta-FAIR","slug":"meta-fair"},"topicExpertise":[{"id":"68494c8f63371e51d83bde6b","name":"Pascale Fung||Anticipation Committee Chair|Advanced AI","slug":"pascale-fung-anticipation-committee-chair-advanced-ai","editions":[{"id":"684951c963371e51d83bdf31","name":"2025","slug":"2025"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]},{"__typename":"Platform_Person","id":"65c55d249e947c438698b226","name":"Rüdiger L. Urbanke","slug":"rudiger-l-urbanke","position":"Professor of Communication and Information Theory","path":"/people/rudiger-l-urbanke","openGraph":{"id":"openGraph_person/rudiger-l-urbanke","title":"Rüdiger L. Urbanke","image":{"id":"image_gesda-22/person/rudiger-l-urbanke_photo__RC3BCdiger_Urbanke_aljmve","url":"https://res.cloudinary.com/shapeable/image/upload/v1668987585/gesda-22/person/rudiger-l-urbanke_photo__RC3BCdiger_Urbanke_aljmve.jpg","url2x":null,"thumbnails":{"card":{"id":"thumbnails-card-file_gesda-22/person/rudiger-l-urbanke_photo__RC3BCdiger_Urbanke_aljmve","url":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_480/v1668987585/gesda-22/person/rudiger-l-urbanke_photo__RC3BCdiger_Urbanke_aljmve.jpg","url2x":"https://res.cloudinary.com/shapeable/image/upload/c_limit,w_960/v1668987585/gesda-22/person/rudiger-l-urbanke_photo__RC3BCdiger_Urbanke_aljmve.jpg"}}}},"organisation":{"id":"65c55d059e947c438698ac3b","name":"EPFL","slug":"epfl"},"topicExpertise":[{"id":"66cbf7840177877dc528d49d","name":"||Anticipation Committee Chair|Advanced AI","slug":"anticipation-committee-chair-advanced-ai","editions":[{"id":"66ab1ba336a8f2f336a557b5","name":"2022","slug":"2022"},{"id":"66ab1bac36a8f2f336a557ba","name":"2023","slug":"2023"},{"id":"66ab1bb636a8f2f336a557bf","name":"2024","slug":"2024"}],"topic":{"id":"65c55d599e947c438698b7a7"}}]}]}}},"pageContext":{"id":"65c55d599e947c438698b7a7","name":"Artificial Intelligence","slug":"artificial-intelligence","entityPath":"/topics/artificial-intelligence","entityTypeName":"Topic","__type":"Topic"}},
    "staticQueryHashes": ["1044227382","1071300789","1158597448","1242646999","1427075558","1452322194","1520036161","1586309863","1606754935","1734835243","1816168740","1903214493","1989845544","2034981229","2097402925","2181044613","239250450","2409034939","2463401854","2612382444","2862279633","2910164142","2912920178","2955142335","3005265507","3073584486","3124191253","3172506128","3320076387","3624873332","364221563","3665990662","3692255024","3778988535","3782604890","4091857177","4216505212","701411134","758936535","881103158"]}