Verified, then logged.
Sun 22 Nov 2026 · 00:22 KST / 15:22 UTC한국어 | ENSearch /

04 / Policy · Policy

Kwon presents GPT-based policy simulation research

The research tests whether GPT-based simulations can predict complex consumer responses, according to DigitalToday.

Shoppers with varied baskets stand at a neighborhood grocery checkout behind a research monitor comparing simulated purchasing responses. AI-generated image

Key points

  1. 01DigitalToday report on 11 Nov 2026 on Kwon's policy simulation presentation
  2. 02Consumer spending coupon effects as an application of large language models
  3. 03Replacement of researcher-set behavioural rules with GPT-based simulations

DigitalToday, a Korean technology news outlet, reported on 11 Nov 2026 that Kwon Hyuk-gu, an endowed professor of information systems at Singapore's Nanyang Technological University (NTU), presented research on GPT-based policy simulation at a media session that day. The work is at the research stage, with its predictive capabilities still being tested.

The session on AI use in policy research took place at OpenAI Korea, according to DigitalToday. OpenAI supported the research into incorporating generative AI into existing rule-based statistical forecasting.

Research approach. DigitalToday reported that researchers traditionally set individual rules governing behaviour in these models. The research replaces those rules with simulations based on OpenAI's GPT models.

Policy application. DigitalToday described research on the effects of consumer spending coupons as an application of large language models (LLMs) to policy simulation. LLMs are AI models trained to process and generate language.

According to DigitalToday, whether GPT-based simulation can predict complex consumer responses remains a question for the research to test.