A coalition of civil society groups, trade unions, academics, and campaigners has sent an open letter to the science and technology secretary, Peter Kyle, calling on the government to abandon plans that would remove the right to avoid decisions made solely by automated or AI systems.

The signatories of the letter include prominent human and digital rights organisations such as Amnesty International UK, the Open Rights Group, and Privacy International, along with others from across the tech sector, including the Scottish Law and Innovation Network, Connected by Data, and Foxglove, a non-profit that works to keep tech firms fair and accountable. 

Other signatories describe themselves as working across fields impacted by automated decisions, including data policy, racial justice, employment, health, disability rights, welfare and criminal justice.  

They argue that automated decisions are sometimes incorrect, with the potential for machines to introduce bias and for AI models to be based on incorrect, uncontextualized data, saying: “The potential for discrimination and for deepening inequalities is known and significant.

“Important machine decisions can be wrong and unjust, and frequently Artificial Intelligence providers are unwilling or unable to address shortcomings. Where we hope both we and the government can find consensus is that Artificial Intelligence technologies, in order to succeed, need to maintain the confidence of the public. The public must be able to trust the way that these technologies are employed.”

Rather than advocate for automated decision making to be scrapped entirely, the coalition is arguing that the government should re-examine changes set to be introduced under the Data (Use and Access) Bill, legislation that the new government hopes can unlock greater use of data, driving innovation and economic growth.

The provisions in the DUA Bill will allow automated decision-making in ‘low-risk scenarios’, while maintaining specific protections for sensitive data, with the aim of easing restrictions imposed by Article 22 of the UK GDPR, which limits the types of data that can be used as well as its intended end use in automated decisions. 

The Open Rights Group claim that if the DUA Bill is to pass, individuals may lose the right not to have AI or other systems make legal, binding decisions without human oversight, although it will still be possible for people to challenge these and request human review when decisions significantly affect them, according to techUK.

Recommended reading

• Report: 92% of ICT Jobs Need Reskilling as Technology Roles Evolve
• AI Skills Gap Puts Public Sector Projects in Jeopardy
• As the AI Revolution Marches On, Who Could Be Left Behind?

“Many AI systems have been proven to magnify discrimination and inequality in society,” said James Baker, platform power programme manager at Open Rights Group.

“We currently have the right not to be subject to a legal, or significant decision based solely on automated processing. It’s vital that the government reconsiders these proposals and protects us from harmful decisions.”

Another piece of legislation, the Public Authority Algorithmic and Automated Decision-Making Systems Bill, is currently scheduled for a second reading on the 13th of December in the House of Lords, where it originated. 

The aim of that Bill is to ensure that automated decision making systems, including algorithmic ones, are ‘deployed in a manner that accounts for and mitigates risks to individuals, public authorities, groups and society as a whole’, and if passed would require public authorities to adopt impact assessments and transparency standards when using such tools.