One way to do data ethically
Data ethics can seem like a nebulous and vague concept. It has come to the fore as a consequence of severe misuse and absue of personal data by previously trusted companies. Helen Lippell, a taxonomy, metadata and search consultant, worked on a project in which ethics were a tangible part of the process.
Here is what Lippell said about how it was done:
“I worked at SSIX, an EU-funded research project involving sentiment analysis technology on social media and public data, such as Twitter, blogs and newsfeeds.
“The idea of the project was to build tools for SMEs within the EU that would be more cost-effective and ethical than the commercially-produced ones.
"We got bashed over the head about privacy."
“Because the project started just before GDPR came in, there was a really strong focus from the funding bodies of having to think about privacy. We got really bashed over the head about it, which was good.
“The way we dealt with it was to set up a completely separate data ethics board which was completely separate to the project team. I was part of that as one of a team of external advisors from media and technology.
“We sat outside the day-to-day project, but we reviewed the documentation. We made recommendations about how to do things better because the thing about Twitter is you can pretty much hoover up whatever you want without consequences. But we didn't want to do that. We wanted to be clear about what we taking and what we were not taking.
"We needed to give people a choice if they didn't want to take part."
“What we needed to do was to give people a choice if they didn't want to be a part of it. As part of that ethical process, we developed a particular tool to allow individuals to say, ‘no, I don't want my data to be part of this project’ in a really easy form.
“Then their details would go on to a blacklist and that will stop any of their stuff appearing in the products that we were building.
"The ethics and privacy element was key."
“That was really useful because we were analysing sentiments and people's opinions. We were making judgements about what people were saying in order to try and make predictions about financial markets. So clearly the ethics and privacy element was quite key.”
Data ethics is always going to be a tough issue to deal with due to the subjectivity of what is ethical. As Inma Martinez said: “Your good may be my bad.”
However, it seemed in this case that the ethical way to deal with people’s data was to give the data subjects a clear and easy-to-execute opt-out and to appoint a panel of overseers to make sure that those carrying out the project stayed on the straight and narrow.
Helen Lippell was speaking at BCS Data Management Specialist Group at the Chartered Institute for IT.