Trade Union Congress draft Artificial Intelligence (Regulation and Employment Rights) Bill

The Trades Union Congress (“TUC”) recently published its proposed Artificial Intelligence (Regulation and Employment Rights) Bill (the “Bill”), which aims to regulate employers’ use of AI as it relates to the rights and interests of workers. 

The provisions detailed in the Bill largely concern rights and obligations which apply where an employer uses AI systems as part of ‘high-risk decision-making’ processes. ‘High-risk decision-making’ refers to decisions which have the potential to have legal, or otherwise significant consequences for employees, workers, or jobseekers. 

Many companies operating in the gambling and games industries will have already implemented AI systems into their internal processes or, if not, will be in the process of doing so. Indeed, lots of online gambling products currently on the market incorporate AI in some form. For example, automated detection and flagging of suspicious gaming patterns that suggest a player is cheating. The existing reliance on AI in these industries makes its proposed regulation a very relevant topic for gambling and games companies.

For instance, the Bill:

• imposes obligations around consultation with trade unions when high-risk decision-making systems are proposed, alongside ongoing annual reviews; 
  
• requires employers to share with trade unions any data collected about union members; 
  
• imposes a duty on employers, and their agents, to perform a “Workplace AI Risk Assessment” before undertaking any such high-risk activity. This assessment requires a consideration of health and safety, human rights, equality, and data protection concerns; and 
  
• requires employers to consult with workers before deploying AI systems, be transparent about the use of such systems in decision-making processes, maintain a register of any such systems which are put in place, and provide a personalised explanation to employees of any such decisions made in relation to them.

The Bill outlines several explicit prohibitions, including a ban against relying on high-risk applications of AI in decisions around employee dismissal, and restrictions on the use of emotion recognition technology where it could detrimentally impact workers.

It also prohibits discrimination through the use of AI and amends the Equality Act 2010 to shift the burden of proof onto employers to show that no discrimination happened in AI or human-led decisions.  Employers would have a statutory audit defence to fall back on, meaning they would not be liable for the discriminatory consequences of AI systems if they can show that (i) they did not create or modify the systems and (ii) they conducted thorough audits and introduced procedural safeguards before deploying them. 

Finally, the Bill also proposes new rights such as the right to disconnect (the right to be unavailable to employers outside working hours), and protection from dismissal or detrimental treatment for exercising that right.

It is perhaps unsurprising that the TUC’s Bill comprises concepts imported from the European Union. For example, definitions around high-risk decision-making closely mirror those outlined by the upcoming EU Artificial Intelligence Act, and the European Parliament has called on the European Commission to draft a new EU law granting a right to disconnect, already part of legislation in several countries across Europe. 

Importantly, the proposals are a Bill only and may never be adopted into Law, but it is an interesting and somewhat divisive topic for the workplace generally. It remains open to future UK governments to consider the TUC’s proposals which may mean we eventually see them come into force. Indeed, in light of the upcoming UK general election, a Labour government may be more amenable to implementing policies in support of Trade Unions and there are clear overlaps with the party’s election manifesto (for example, championing the right to disconnect).