View from Westminster Sign up for emails to get expert analysis delivered straight to your inbox.
Get your free View from Westminster email
EEveryone agrees that AI is already doing a lot of good. For example, mammogram analysis will be faster and more reliable.
This technology is advancing at the speed of light and is sure to bring even greater achievements in the future. Clearly, as a nation, we need to make the most of our many assets. However, the government's response to this new technological frontier appears to be small-scale.
We need an AI workforce strategy that is as urgent and long-term-focused as the NHS workforce strategy. We need to ensure that AI is part of our leveling up agenda, widening its opportunities and leveraging talent from all regions, including Scotland and Wales.
We are working individually, through trial and error, although we need the coordination of all public sector organizations looking for ways to move forward on this issue. And we need to see data as part of our infrastructure development, as important as the roads and rail lines we use to commute.
To ensure a smooth development trajectory for this technology here in the UK, we need strong regulation that enables AI to take off, but also identifies and prevents problems before they occur.
The usual pattern of regulation is that primary legislation is supplemented by secondary legislation, passed by Parliament and enforced by regulatory authorities empowered by law.
But the problem with AI is that even as it moves at breakneck speed, the pace of legislation is painfully slow. Especially when Congress is trying to wrap its head around something new and complex. What happens is that our laws are already years out of date by the time they receive Royal Assent.
It took 10 years from Prime Minister David Cameron's initial announcement to the passing of the Online Safety Bill. This is a desperate timescale for AI regulation. You can't enact meaningful regulatory laws if you can't even fully imagine what you're regulating.And we know it will happen, but we can't fully predict it how AI will develop over time.
The answer is not to leave everything to the government and regulate it through executive authority. These are important public policy issues, and Congress must do its part. Policy decisions cannot be left to regulators. Congress must decide public policy, and the role of regulators must be to implement it.
Through Congress, we can get fast and smart laws and regulations that are professional and real-time. But this will require changes to parliamentary processes to address this challenge. We could strengthen the Science, Innovation and Technology Committee (perhaps jointly with the Business and Trade Select Committee) with additional resources and give the Select Committee regulatory powers on behalf of Congress through primary legislation. You can do it.
The bill would give the Select Board broad regulatory powers. Parliament will then continue to consider the work of the Select Committee in the normal manner. That would create a committee of legislators with real expertise and regulatory authority built into Congress to keep up with the pace of AI development.
There is plenty of precedent for Congress adopting new processes when necessary. New processes for regulating AI are a prime example of change.
The problem that governments are not recognizing is the risk that AI will become the domain of a “tech bro” culture, embedding bias and discrimination in data sets and thereby perpetuating the discrimination that it is public policy to eradicate. Discrimination on the basis of gender, race, or social class is bad enough if it can be explicitly recognized and challenged. But hidden biases that cannot be reversed can be frightening. If it contaminates our datasets and gets written into our algorithms, it could halt the progress we have made toward equality and even turn back the clock. Among the many organizations working in this space, the Open Data Institute (ODI) advocates a commitment to strong data infrastructure as a critical step to tackling inequality, bias, and discrimination.
There is already an annoying “tech buddy” culture in AI. Research conducted by the Fawcett Institute earlier this year found that almost half of men working in the industry believe they do not benefit from gender balance in the workforce, and one in five believe that women are less likely to benefit from gender balance in the workforce. It turns out that he doesn't think he's very suitable for the job. Studies have shown that this attitude prevents women from staying in the industry, resulting in half the workforce being lost to women.
We all have a stake in the advancement of AI. We all need a say through Congress in how it is regulated, and we need women as well as men to drive it.
Harriet Harman is the Labor MP for Camberwell and Peckham.
