Open Data Institute Podcasts

Open Data Institute Podcasts is the feed of the Open Data Institute, a non-profit based in London. Its recent episodes fall into three formats, and knowing which one a recording belongs to says a lot about what it will offer. The first is the workshop session. 'Data Ecosystem Mapping (Spanish Edit)' sets out to explain data ecosystems and data infrastructure before turning to the mapping tool the Institute developed. It grounds the idea in a plain business example, a network of manufacturers, suppliers and distributors, and makes the case for open standards as reusable agreements that help people access, share and use good-quality data. Its closing argument is that value and impact should be judged across a whole ecosystem, not just inside one organisation. 'Data Ethics Canvas (Spanish edit)' is led by Lucas Stiglic, a public policy associate at the Institute, who presents the Canvas as a way to identify and address the ethical questions a data project raises. One passage separates law from ethics with real care. Law is written down and backed by the state, while ethics can live in codes of conduct, professional bodies and unwritten social norms. That session begins in Spanish. The second format is the interview. Stuart Coleman, the Institute's Learning and Business Development Director, describes Inside Business as a discussion between himself and two or three data leaders. 'How access to real time national data could help businesses recover from the pandemic' departs from that with a one-to-one with Fiona James, Director of Data Growth and Operations at the Office for National Statistics. Their conversation covers pre-linked government data, secure access for accredited researchers that can include businesses, and a service directed partly by user need. Coleman also raises the checks organisations might face to show they work in the public interest. The third format is the provocation. In 'Provocation 1 – Juan Mateos-Garcia, Nesta', the speaker argues that modern AI is a technological experiment that might prevent future experiments, which matters because experimenting is how people learn and change course. He then lists the externalities: models that reproduce bias against disadvantaged groups, labour displacement, and bad actors generating misinformation or toxic content. He links these to questions of openness and closes on policies that keep the benefits while mitigating the risks. For a first listen, the Data Ethics Canvas session offers the most usable framework, while the provocation suits someone who would rather be argued with.