Sandra Wachter

Why organisations work with Sandra Wachter

• Her Counterfactual Explanations framework – co-developed at Oxford and adopted in production by Google, IBM, Microsoft, Accenture, and Vodafone – gives organisations a legally defensible method for explaining algorithmic decisions to the individuals they affect, not just to internal audit teams.
• Her Conditional Demographic Disparity (CDD) bias test demonstrated that 13 of 20 widely used bias detection tools fail EU non-discrimination standards. Amazon and IBM implemented CDD in their cloud services; it was deployed in 2024 to expose systemic bias in the Dutch national education system, resulting in a ministerial apology and formal reform.
• She works at the precise intersection of legal standards and technical metrics – meaning she can tell risk and compliance teams not only what accountability requires but how to measure whether systems actually deliver it.
• As a member of the European Parliament Working Group on AI Liability and the World Bank Task Force on Access to Justice and Technology, she tracks how regulatory frameworks are being written in real time, giving organisations visibility into requirements before they become mandatory.
• Her research has shifted institutional practice at the NHS and MHRA, which are revising medical device licensing standards using her findings on the unintended consequences of common group fairness measures – a concrete precedent for how her work moves from publication to policy.

Biography highlights

• Professor of Technology and Regulation, Oxford Internet Institute, University of Oxford; Humboldt Professor of Technology and Regulation, Hasso Plattner Institute
• Leads the Governance of Emerging Technologies (GET) Research Programme at Oxford
• Co-developer of Counterfactual Explanations (now deployed by Google, IBM, Microsoft, Accenture, and Vodafone) and the Conditional Demographic Disparity (CDD) bias test (implemented by Amazon and IBM; applied in the Netherlands education reform, 2024)
• Published in Science Robotics, Nature Electronics, and Nature Reviews Physics; cited over 20,000 times (Google Scholar)
• Alexander von Humboldt Foundation Research Award (2025, €3.5M); O2RB Excellence in Impact Award (2018, 2021); Computer Weekly Women in UK Tech Award (2021); Privacy Law Scholar Award (2019); CognitionX AI Ethics Award (2017, 2023)
• Policy roles: European Parliament Working Group on AI Liability; World Bank Task Force on Access to Justice and Technology; WEF Global Futures Council on Values, Ethics and Innovation; UNESCO; UK Police Ethics Guidance Group
• Previously Visiting Professor, Harvard Law School; Research Fellow, The Alan Turing Institute
• Media profiles in the Financial Times, Wired, New York Times, BBC,Guardian, Harvard Business Review, Reuters, Time Magazine, and MIT Technology Review

Biography

Sandra Wachter is Professor of Technology and Regulation at the University of Oxford’s Oxford Internet Institute and Humboldt Professor at the Hasso Plattner Institute. She leads Oxford’s Governance of Emerging Technologies Research Programme – work that sits not in the space of ethical aspiration but in the harder territory of legal standards and technical verification.

The tools her research has produced are in use at scale. The Counterfactual Explanations framework she co-developed is deployed in production by Google, IBM, Microsoft, Accenture, and Vodafone. Her Conditional Demographic Disparity (CDD) bias test – built after demonstrating that 13 of 20 commonly used bias tools fail EU non-discrimination law – was adopted by Amazon and IBM in their cloud services. In 2024, CDD was used to expose systemic bias in the Dutch national education system; the Dutch Minister of Education apologised and initiated formal reform. The NHS and MHRA have revised their medical device licensing practices using her findings on the consequences of widely applied group fairness measures.

Her publications appear in Science Robotics, Nature Electronics, and Nature Reviews Physics, and she has been cited over 20,000 times. Awards include the Alexander von Humboldt Foundation Research Award (2025, €3.5M), the O2RB Excellence in Impact Award (2018, 2021), the Privacy Law Scholar Award (2019), and the CognitionX AI Ethics Award (2017, 2023). She sits on the European Parliament Working Group on AI Liability, the World Bank Task Force on Access to Justice and Technology, the WEF Global Futures Council on Values, Ethics and Innovation, and the UK Police Ethics Guidance Group, and was previously Visiting Professor at Harvard Law School.

For boards and risk leaders, her most immediate value is diagnostic: the gap between what organisations say their AI governance achieves and what the regulatory and technical evidence supports is larger than most assume. She has spent a decade building the instruments to measure that gap precisely.

Key speaking topics

• AI governance and regulation
• Algorithmic accountability and explainability
• Algorithmic bias and non-discrimination law
• Generative AI and legal liability
• The EU AI Act and regulatory compliance
• Profiling, inferential analytics, and data protection
• Platform regulation and internet governance

Ideal for

• Boards and C-suite teams making AI adoption and governance commitments
• Chief Data Officers and technology leadership teams designing accountability frameworks
• Legal, compliance, and risk leaders responsible for AI deployment decisions
• Public sector and regulatory audiences shaping AI policy and standards

Audience outcomes

• A clear-eyed assessment of where current AI governance practices fall short of EU and UK legal standards – and why common assumptions about bias testing are insufficient
• Practical grounding in legally defensible tools for explainability and bias measurement, including frameworks already adopted by major technology companies
• Informed foresight into how the EU AI Act and emerging AI liability frameworks will affect deployment requirements
• Understanding of how generative AI creates specific new liabilities around hallucination, misinformation, and the emerging legal question of AI truthfulness
• The ability to distinguish between stated accountability commitments and governance that can withstand external scrutiny