AI advances could help empower the financially excluded but it must be better regulated says shadow industrial strategy minister
Advances in AI could help empower the financially excluded, but new technologies need to be better regulated, shadow industrial strategy minister Chi Onwurah has said.
Speaking to a packed fringe event on the ethics of AI in the financial services, Chi Onwurah welcomed the benefits of technology advances for customers but warned the increasing complexity of the systems were bringing new challenges for the sector.
“There is huge potential in reaching out to the financially excluded, not through technology on its own, and certainly not using AI without appropriate transparency and regulation”, she said.
“The opportunity with algorithms is to personalise and customise the experience of everyone. But the challenge is that algorithms also automate and industrialise bias. And the bias, it may be in how the algorithms are designed, but it is often actually in the training data.”
And Ms Onwurah called on the government to ditch its “light touch” approach to regulation, instead urging tech companies to prioritise transparency by providing regulators with access to their algorithms to ensure that firms were responsible for the long-term impacts of their work.
“Just as open source software has ensured we have software that is not only cheaper and more available but which is also improved and de-bugged as it is continuously tested, then there is an argument for at least sharing the principles or some of the workings behind some of these algorithms so that we can be sure that they are working to help and empower people in accessing financial services, and not to impose decision making upon them.”
The panel, organised by the Chartered Body Alliance, a joint initiative by the Chartered Insurance Institute, Chartered Institute for Securities & Investment and the Chartered Banker Institute, focussed on how AI was changing the landscape of financial services with a spotlight on how fairness and equality could be built into new and advanced systems.
Dr Adrian Weller, the AI programme director at the Alan Turing Institute, highlighted how the increasing use of specialised algorithms had already delivered a series of benefits for consumers.
He said: [They have brought] great improvements in fraud detection, trade processing, compliance, and where appropriate, personalised recommendations of products, and for insurance personalised pricing depending on individual risks.
But Dr Weller also flagged that the growth of these large interconnected systems could also multiply risk when they fail: “We are seeing greater interconnectedness of systems, and that tends to mean things work really well typically , but when things go wrong, they can potentially go wrong in worse ways.”
Sian Fisher, CEO of the Chartered Insurance Institute, also welcomed the new possibilities afforded to consumers by the adoption of these new technologies by the financial services sector but said that companies had to ensure they were “not just getting carried away and really excited about building the tools, but actually taking responsibility for the outcomes.”
She added: “I thought it was quite interesting hearing Mark Zuckerberg – whether he wanted to or not – admitting that that was probably one of the things they had missed slightly, that you can’t create the tool and stand to the side of it. If you create the tool, then you’ve also got to take responsibility for the outcomes.”
And Dr Weller agreed that companies should be held accountable for their use of AI and machine learning, adding: “If an algorithm is doing something, we need to make sure that whoever is behind it is held accountable if the algorithm goes wrong. It’s a really important principle we need to keep in mind.”
Ilford North MP Wes Streeting argued the best way to mitigate the risk of future technologies was to encourage technologists to focus on creating technology that was based of principles of fairness and democracy.
“There is a big revolution taking place technologically which throws up amazing opportunities for society, for businesses, for nations, for the whole world, but also poses some significant risks”, he said.
“There are some significant risks, particularly around making sure that at the heart of technological change, you have principles which are democratic, fair, focussed on reducing human biases and tackling vesting interests.
“There are two futures available, and we need to mitigate against some of those risks, for example particularly around data and the ubiquity of data, the utility of data.
He added: “Data is going to become a more powerful commodity arguably, but we have got to make sure as much of that power, wealth and opportunity rests with people who should have greater ownership over their data.”