How to stop data from driving government mad

Spare a thought for Dominic Cummings - it was data that drove him mad. Data drives us all mad. For a field to be successful today it must curate, store, and share data across disparate teams, which is maddeningly difficult. How might governments learn from this?

The Institute for Government's Commission on the Centre of Government has revealed how underpowered Number 10 is, not least in data analysis. Its report into reforming the institutions at the centre of government recommends creating a Department for the Prime Minister and Cabinet and, within that, a Joint Analysis and Assessment Centre (JAAC) to assess and analyse data for better decision-making.

This would be a start, but everybody in large organisations knows that top-down initiatives from the centre rarely work well at the coalface. If the JAAC is to be effective at converting data into information, what insight could it glean from structures that have evolved to do this? And what could it learn from scientific fields that manage this successfully?

First, deep neural networks learn by repeatedly passing information back and forth until every neurone is tuned to achieve the same objective. Information flow in both directions is the key.

The free flow of data and information between the centre and departments will be key

Neil Lawrence, DeepMind professor of machine learning at the University of Cambridge, notes that in government, "People at the coal face have a better understanding of the right interventions, although not what the central policy might be; a successful centre will have a co-ordinating function driven by an AI strategy, but will devolve power to the departments, professions, and regulators to implement it."

Or, as Jess Montgomery, director of AI@Cam says: "Getting government data - and AI - ready will require foundational work, for example in data curation and pipeline building." The free flow of data and information between the centre and departments will be key. This is an obvious point, which is far easier for computers than for people, but it must not be forgotten.

Second, science and technology are team sports played with equipment developed over generations. Thousands of years of human endeavour are embedded in a computer chip and in the algorithms that run through it. As this knowledge accumulates, so does the value of the expertise required to navigate it. In science, as in sport, successful teams contain players with different but overlapping skills. No individual can win alone.