New guide launched to encourage ethical approach to data analytics

The ‘Good Data Guide’, unveiled today by data consultancy firm Profusion, is a practical manual designed primarily for small to medium-sized businesses that often lack the resources of larger companies. Profusion said the guide can be used “tactically”, to address specific organisational issues or to evaluate new data initiatives.

Natalie Cramp, CEO of Profusion, said: “Data has changed how we travel, manage our finances, interact with one another, and consume content. Elections have been impacted, infrastructure and agriculture made more efficient, and entire industries transformed. And we are only at the start of the journey.”

“Whether the destination we reach will be for the better or worse of society will be predicated on how we answer the big ethical questions. Is my data biased? Does this algorithm treat all users fairly? Can we explain how our algorithms make decisions? Are our development processes transparent and inclusive? Where are the boundaries of privacy? How do we enable informed consent? How do we ensure there is accountability for how data is used?” Cramp said.

She added: “Our most important objective with the Good Data Guide is to demystify the subject and to make data ethics accessible and achievable for all.”

Dr Sue Chadwick of Pinsent Masons, who specialises in data ethics and the built environment, and sits on Profusion’s data ethics advisory board, warned that it was more important than ever for organisations to take a proactive approach to data transparency and diversity. She said: “Laws are slow to evolve and are constantly in a state of catch up with human innovation and technological developments.”

“Last summer, when grappling with the question of whether a private hire vehicle service run from an online platform should have given an operator’s licence, a judge was forced to rely on the Metropolitan Public Carriage Act of 1869. There is no better example to show why regulation alone is not enough to deal with the impacts of an increasingly digitalised world,” she said.

With lawmakers constantly playing catch-up with technological progress, Chadwick said organisations had to shoulder some of the responsibility. “That means ensuring that everybody in an organisation knows and understands its data ethics principles, as well as making sure that every decision a company takes includes consideration of its data ethics concerns,” she said.

Chadwick explained that organisations also need to consider the makeup of the teams that organisations pick to handle their data. “They need to be careful about the data that is used, understanding where it comes from and ensuring that the underlying datasets represent the population properly,” she said.

“Beyond this, organisations need to understand how specific issues like algorithmic bias affect the use of sensitive data – particularly biometrics – because it carries serious equalities implications. As the Good Data Guide explains, conducting an equalities impact assessment, whether or not one is required by law, can help an organisation ensure it is fulfilling its ethical obligations concerning diversity and data, and this is in line with a recent speech by Lord Sales, a Supreme Court judge,” she said.

“Ultimately, to achieve good data practice an organisation cannot think of data ethics as an ‘add on’– it must be treated as an integral part of its operations. It requires organisations to rewire everything they do, to include considerations of data ethics,” Chadwick said.