A major funder has assured charities it is "alert" to risks around bias and loss of trust as it trials the use of artificial intelligence to assess applications.

The Clothworkers' Foundation is testing how well AI can help assess applications for its Open Grants Programme, which handed out £8.65m in grants last year

The decision has been taken as demand for the funding programme “continues to grow”, said the funder in a LinkedIN post around its use of AI in the application process.

Currently the funder has two members of staff assessing the 1,430 eligible applications it received for the programme this year.

It explains that during the summer it began trialling Plinth AI to see how it can be used to assess applications.

Following these initial tests the funder says it is “confident enough in the results to expand the trial” with Plinth AI becoming one of the two “pairs of eyes” to assess eligible grants.

But it adds that a staff member will also assess applications and “will always have the final say in case of disagreement. AI will not be able to reject any application – it will only assess”.

It added: “After this stage, members of the team, and in some cases Trustees, will continue to decide which applications actually receive funding.

The use of AI in the application process is taking place as it prepares to undertake a strategic review next year into its grantmaking over the next five years.

The funder says it is looking to manage potential risks of introducing AI into its grantmaking, including “importing new biases”.

“We aim to be clear about which organisations and projects we prioritise and it is important that AI assessment and scoring aligns with these and does not advantage or penalise organisations that our team would not,” said the grantmaker.

“Our trial has compared AI judgments with our own past assessments, and we will continue to compare it to our current assessments to maintain consistency. However, we are maintaining human judgment and ultimate decision-making for every application.”

It adds that it is also mindful of “loss of trust from our applicants”, saying “we are committed to being transparent about how and where we are using AI to mitigate against this”.

"In our piloting of Plinth we are alert to these risks and others and we want to understand how well we manage them, as well as how good AI is at assessing grants," it adds.

---