Benefits of ‘welfare robots’ and the need for human oversight

https://www.theguardian.com/society/2019/oct/17/benefits-of-welfare-robots-and-the-need-for-human-oversight

Version 0 of 1.

Re your article (March of the ‘welfare robot’ triggers fears for poorest, 15 October), I found it disappointing and surprising that you do not see the benefits of the DWP’s work in artificial intelligence, despite a recent editorial (22 May)praising the benefits of AI in helping patient care in the NHS.

Our “intelligent automation garage” is using technology to improve the experience for claimants, bringing it in line with the service most of us expect and enjoy from our banking apps, shopping websites or utility providers. We are establishing a welfare system for the future.

Far from being “ruled by computer algorithms”, our bots don’t make decisions regarding people’s benefits. Instead they focus on everyday repetitive tasks so our colleagues can spend more time supporting vulnerable claimants face to face. Our use of machine learning is making the system simpler for people – so, for example, it won’t ask for more information than is necessary for straightforward claims such as childcare and housing. Importantly, digital-first is not the same as digital-only – as the Guardian, a digital-first platform since 2011, will understand.

While we proudly invest in our award-winning digital team, we continue to support vulnerable people in our jobcentres every day and ensure there’s still face-to-face support for those who need it.Simon McKinnonChief digital and information officer, Department for Work and Pensions

• Your article (One in three councils using algorithms to make welfare decisions, 16 October) risks throwing the baby out with the bathwater. There are of course dangers. Predictive algorithms can have negative consequences, but so can human decision-making. The answer lies in careful and considered deployment of new technologies, not dismissing the potential of AI systems out of hand.

We would encourage local authorities to give proper scrutiny to the use of algorithms in decision-making processes. Specific questions, relating to bias in data sets, risks and harm mitigation, ethics and human oversight must be asked. Councils should also strive for greater transparency in how they are using data and establish mechanisms for citizens to be involved with the design process and see which aspects of their data are being used and how.

But to deny the potential for AI to improve public services outright is unhelpful and risks genuine benefits for citizens being missed.Tom SymonsHead of government innovation, Nesta

• The error rate of the current computerised benefits payment system causes terrible suffering to those who are most at risk when it goes wrong. People without money can’t buy food, and many can’t get to a food bank as it could be miles away. Having to hook up on a computer at a local library is a step too far for thousands – assuming the library hasn’t closed down. We must have consultation on AI and the benefits system, and halt this steady move towards machines ruling our lives. Pat McCarthyBrentford, London

• Join the debate – email guardian.letters@theguardian.com

• Read more Guardian letters – click here to visit gu.com/letters

• Do you have a photo you’d like to share with Guardian readers? Click here to upload it and we’ll publish the best submissions in the letters spread of our print edition