Rethinking data power: beyond AI hype and corporate ethics

As artificial intelligence systems gain unprecedented power, the corporate “data ethics” movement has all but collapsed, leaving a critical gap in governance. At a recent event on “The power of data: ethics, politics, and public interest“ co-hosted by LSE’s Data Science Institute and the Department of Media and Communications, chair Seeta Peña Gangadharan brought together three leading experts to examine how data power and AI systems are reshaping society, often without meaningful public oversight.

Data must be understood within a nexus of power and truth

, Columbia University professor and chief data scientist at the New York Times, opened the discussion by tracing the evolution of data from Victorian-era statistics through wartime computation to today’s AI boom. His core argument was that data must be understood within a nexus of power and truth. He emphasised that data is never neutral; it reflects the assumptions, interests, and biases of those who design and deploy it. Wiggins offered a historical lens, noting how early statistical methods were shaped by their creators’ biases. Citing simultaneous promotion of correlation, regression, and eugenics, he showed how science has always been entangled with ideology. Quoting , Wiggins challenged the audience to question the belief that today’s science is free from such prejudices.

This entanglement continued into the mid-20th century with the emergence of digital computation. As data work became institutionalised, it was immediately gendered. Wiggins pointed out how divisions between men’s and women’s labour in early computing reflected broader social hierarchies, reinforcing the argument that power is embedded in data practices from the start.

Perhaps the evening’s most striking insight was the recent collapse of corporate data ethics initiatives. Between 2017 and 2023, there was a flourishing of literature exposing algorithmic harms, such as Weapons of Math Destruction and work on racial bias in search engines. Yet by 2023, as AI technologies expanded rapidly, tech companies were laying off ethics teams. Wiggins explained that “ethics” became a “drift word” appropriated and redefined by corporate interests. He cited law professor frustration: “What I always see in the AI literature these days is ethics. I want to strangle ethics.” The result was a hollowing out of critical ethical frameworks in favour of corporate-friendly narratives.

In the panel discussion that followed, Wiggins mapped out what he termed a “three-player unstable game” between corporate power, state power, and people power. He emphasised that regulation does not only consist of legal frameworks but also includes norms, market forces, and technical architectures. People power, he argued, can be mobilised through collective action by tech workers, whistleblowers, and consumers. For instance, consumer boycotts can significantly influence corporate behaviour and long-term market dynamics.

Data collection always involves storytelling

, Associate Professor at LSE and Director of the Data and Society programme, shifted focus to the stories we tell about data. She posed three questions: Just AI Project. The method begins with a refusal to accept pre-given definitions of data. Instead, participants are encouraged to debate the meaning of data, as these debates shape what is considered important in public life. Her community work in East Walworth, South London, demonstrated this in action. At a 2023 community festival, local residents discussed collective electricity use in tower blocks, showing how citizens can reimagine data for practical and cooperative purposes.