Council warns residents of AI complaint risks – but could use same technology to send replies

Newport residents who complain to the city council should be wary of using generative AI, according to a new policy – which also shows the local authority could use the same technology to write replies.

The council has set out new rules for dealing with complaints, compliments and comments – but a report warns AI-generated content “can sometimes contain incorrect, incomplete, or unsupported information”.

Generative AI’s popularity has increased significantly, with a majority (58%) of Britons surveyed by pollsters Ipsos earlier this year having used the technology, and around one in three (35%) using AI tools on a weekly basis.

Using AI is “a personal choice” for each resident, but the council said anyone getting in touch is responsible for making sure the information they provide is “accurate, relevant and reflects their concerns”.

The council has asked residents to “review any AI-assisted content carefully before submission”, and warned “reliance on inaccurate information may affect our ability to understand and respond to the issues raised”.

But the same policy shows the council may itself use AI to support “limited administrative aspects” of the complaints process, including “drafting correspondence”.

In the policy, the council said the complaints it receives are “always assessed, investigated and determined” by officers.

It adds AI is not used to make decisions about complaint outcomes or remedies – and “responsibility and accountability for all responses remain with the council”.

Officers will review content after AI is used, and the technology will not be deployed for “automated” decision-making.

The report adds the council is “committed to openness and fairness in its complaints handling and will always ensure that the use of technology supports – rather than replaces – effective human oversight”.