What will the General Election mean for AI regulation?

Artificial intelligence will transform the world of work in the next decade and will have major economic and social implications across the globe.   So it's surprising (for which read: not at all surprising) that so far it has barely featured in the General Election campaign, with the parties saying very little about their plans.   However, although policy detail remains thin on the ground, there seems to be a clear divergence of approach between the Conservatives and Labour. 

The Conservative Government's approach to AI development thus far has been criticised as overly passive and laissez-faire.  Whether that's fair is open to debate, but it's certainly true that their approach has been to hold back on legislating - they seem in no hurry to introduce a UK equivalent of the EU AI Act (which itself was watered down significantly during its passage).   The ill-fated Digital Information and Data Protection Bill has been one of the casualties of the election period, as it wasn't included in the legislation “wash-up” before the election purdah period.   

The Government previously identified 5 key principles of AI regulation which it expects existing regulators (such as the FCA) to implement: 

1. Safety, security, and robustness
2.  Transparency and explainability
3. Fairness: Designing and using AI systems in a way that avoids unfair bias or discrimination and promotes equal treatment and opportunities 
4. Accountability and governance: Establishing clear lines of responsibility and accountability for AI systems 
5. Contestability and redress

However, in the absence of specific legislation (and without additional funding to hire staff with the requisite expertise) this framework has been criticised by many as inadequate.  Despite the deregulation agenda pursued by the Government, it seems likely that, if they do defy the current odds and win the election, they will need to look again at legislating in due course. 

The signals from Labour are harder to interpret.   They were among the harshest critics of the Government's White Paper on AI regulation, and will be alive to trade union concerns about transparency in the workplace and job displacement, but equally have been keen to present themselves as aligned with business interests and technological innovation.   

Labour's employment policy paper “Labour's Plan to Make Work Pay”, states that the party would consult about how to promote best practice in relation to “the invasion of privacy through surveillance technology, spyware and discriminatory algorithmic decision making”.  This rather gives the impression that Labour would use soft regulation such as ACAS guidance instead of legislation. It also states that Labour would require employers to consult with trade unions or elected staff representatives before introducing new surveillance technologies.  This will be welcomed by trade unions, but doesn't address more fundamental concerns about transparency and redress when AI is used in the employment context.  

AI is complex, controversial and difficult to reduce to a three word slogan, so perhaps it's unsurprising that the main parties have largely side-stepped it.   But one hopes that whoever wakes up as Prime Minister on 5 July is willing to grasp the (virtual) nettle of this technological revolution.