Issue - meetings
South and East Lincolnshire Councils Partnership Artificial Intelligence Policy
Meeting: 12/12/2024 - Cabinet (Item 52)
52 Artificial Intelligence Policy
PDF 127 KB
(A report by James Gilbert, Assistant Director – Corporate)
Additional documents:
Minutes:
Councillor Sandeep Ghosh presented the Artificial Intelligence Policy report, a partnership policy (Appendix A), that provided a framework for how the Council would use AI and established clear accountability mechanisms for managing associated risk. The Policy also explained the concept and types of AI. The policy had been drafted in consultation with Portfolio Holders, Senior leadership Team, Transformation, ICT, the Monitoring Officer and the Data Protection Team.
Members discussed the report and commented in relation to concerns raised regarding potential AI’s generation of bias, discrimination or offensive comments. Members were advised that as AI was a result of a continuous learning process, accuracy would increase with training. Additionally, it was highlighted that the risk would be addressed with intervention procedures.
The recommendations were moved by Councillor Sandeep Ghosh and seconded by Councillor Emma Cresswell.
RESOLVED:
That the Artificial Intelligence Policy, attached to the report as Appendix 1, be approved.
[The Assistant Director – Corporate left the meeting at 7.26pm following consideration of the above item.]
Meeting: 25/07/2024 - Overview & Scrutiny - Corporate & Community Committee (Item 107)
107 South and East Lincolnshire Councils Partnership Artificial Intelligence Policy
PDF 166 KB
A report by the Assistant Director - Corporate
Additional documents:
Minutes:
The Business Intelligence and Change Manager presented the report supported by the Group Manager for Insights and Transformation. Committee were advised that the policy for the partnership had been drafted in consultation with Portfolio Holders, Senior leadership team, the Transformation department, ICT, the Monitoring Officer, and the Data Protection team.
This policy would ensure that AI applications would be aligned with ethical standards, regulatory requirements, and public interest. It would provide a framework for transparency, accountability, and fairness in AI deployment, addressing concerns related to privacy, bias, and security.
By proactively establishing an AI policy across the partnership, its councils could foster innovation while safeguarding community trust, ultimately enhancing service delivery and decision-making processes for the benefit of all citizens. The policy would further benefit the partnership with a consistency across the Partnership workforces which was important as officer teams worked more closely together. In the future the Councils might look to also source single ICT systems to support the delivery of services which common ICT policies facilitated.
Committee deliberation followed which included:
Responding to a question on how much AI was currently in use, a member was advised that it was minimal but was starting to creep into the systems, hence the need for the policy. There were minor AI suites within Teams and PSPS also used some within their services, with a minimal use within the Council.
Voicing concern at the potential drawbacks of AI, a member questioned the lack of AI being able to assess any personality traits. It would not identify any tone of voice and not be able to differentiate, as such it would treat each person in the same way. Committee were advised that there had to be a degree of human oversight. AI used language-based learning and could identify when language became negative or frustrative and could show some form of empathy. Anything generated via AI would need human overview to cover any concerns. Members were further reassured that whilst there was opinion that AI would replace jobs, it was not about replacing jobs. What it would do would be to improve efficiency and as a result, enable vulnerable residents to access operators in person a lot quicker.
On questioning what safeguards were in place to avoid any problems for the Council, officers advised that if a query was not answered in the first one or two attempts, Chatbot would automatically push the call out to a human traditional channel, and there would be rigorous testing across all capacities before any implementation.
Summarising the Chairman noted that the report had been well written and that he agreed the concerns noted that a more
constant review would be required during the first five years. The Business Intelligence and Change Manager confirmed that he would amend the review cycle in line with the request of the committee.