DAEJEON — Artificial intelligence will take part in reviewing and evaluating public procurement to support faster and more reasonable assessments.
The Public Procurement Service (PPS) said on the 26th that it will begin pilot testing an AI-based technical evaluation support system, currently under development, in reviews for Excellent Product certification and in evaluations of negotiated contract proposals. The agency said the goal is to produce more objective and valid assessment results.
The system uses AI to conduct in-depth analysis of vast technical documents, such as the specifications submitted for Excellent Product certification and negotiated contract proposals, providing key information to help evaluators make faster and more accurate decisions.
In Excellent Product certification reviews, the system uses AI to analyze the specifications of recently certified products in the same category and produces a chart comparing the technology and performance of the applicant product against existing certified products. Evaluators can see differences in technology and performance at a glance and can also use a smart search service to quickly find the technical information and related content they need.
For negotiated contract proposals, the AI extracts the main technologies and requirements from the request for proposals and then analyzes the bid proposals. During evaluation, the system will also provide a function that shows the latest technology trends in the relevant field, along with an evaluation-support chatbot service, so that evaluators can efficiently review the core content of lengthy proposals.
The PPS plans to first provide AI-generated technology and performance comparison charts to evaluators on a trial basis for 11 product categories in this year's third round of Excellent Product certification reviews, running from Aug. 26 to Sept. 18. It will then provide AI analysis of bid proposals for at least five evaluations of negotiated contract proposals for IT projects during September, verifying the system's accuracy and usefulness in actual evaluation settings.
The PPS said it has also put safeguards in place to minimize AI hallucination and analysis errors that could arise as the system is introduced. It has built in a source-location feature so that evaluators can directly check and verify the basis for the AI's analysis, and it plans to clearly label all analysis materials as "supplementary reference for evaluation" so that AI results do not influence evaluators' independent judgment.
"We will continue to improve the system's functions by reflecting the improvements identified during the pilot, and we plan to gradually expand its scope to other areas of public procurement review and evaluation, such as innovative product certification," said Kim Ji-wook, head of the PPS's Digital and Fair Procurement Bureau.
