- 630 research teams took part, and the final “AI Champion” was selected to receive follow-up research funding of up to KRW 3 billion

- Top five teams will receive up to KRW 5.8 billion in total, and 15 additional outstanding teams will receive AI innovation research grants

- Technologies recognized in fields such as mobile AI agents, AI-based visual and recycling inspection, and noise-control physical AI

The Ministry of Science and ICT (MSIT, Deputy Prime Minister and Minister Bae Kyunghoon) held the final evaluation and awards ceremony of the 2025 AI Champion Competition on November 5 (Wednesday) at the Dragon City Hotel in Yongsan, Seoul, and announced the five top-winning research teams.

※The following is a chronological list of major press releases issued in relation to the 2025 AI Champion Competition:

“AI Innovators Begin Their R&D Survival — Announcement of the AI Champion Competition” (June 18), “2025 AI Champion Competition Team Registration Results” (July 24), “100 Challenges Toward Innovation Begin — The Stage for AI Champions Opens” (August 11), and “20 Leading AI Research Teams Pass Interim Evaluation of the 2025 AI Champion Competition” (October 20)

Summary of Award Results

Rank Award Title Winning Team

1st AI Champion AutoPhone

2nd AI Innovator Stratio Korea

3rd–5th AI Challenger Revivo / SafeAngel / FineHost

Popular Vote AI Impactor SafeAngel

 Number of research teams in final stages: 20 → 5 winning teams

 Total researchers: 160 → 54

 Average team size: 8 → 10.8

 Fields: Large Language Model (LLM)-based Agents, visual detection, and industrial safety physical AI

 Outcomes: 3 software-based, 2 hardware-based

Despite being held for the first time this year, the AI Champion Competition received enthusiastic participation, attracting 630 teams. Teams formed freely around open AI research topics and advanced through a multi-stage process — preliminary (100 teams), interim (20 teams), and first-round final (shortlisted to 8 teams) — culminating in the final evaluation to determine the AI Champion after several months of intensive competition.

The final evaluation was conducted by a panel of academic, industrial, and research experts, as well as professionals specializing in the market applicability of AI technologies. In addition to the six criteria assessed in the interim evaluation—innovation, practicality, excellence, execution capacity, feasibility, and challenge spirit—the final review also considered growth potential and global (overseas) expansion.

Given that winners receive large-scale follow-up research funding, the final evaluation proceeded in two stages to ensure a thorough review.

The first evaluation round of the final stage (October 31, at the EL Tower in Yangjae, Seoul) required the 20 teams that passed the interim evaluation to exhibit and demonstrate their AI solutions before the judges, who assessed each team’s performance on the spot. Eight teams with the highest scores advanced to the final round.

The final round of evaluation featured presentations from the eight finalist teams before the judges and attendees, showcasing their research achievements. Each team’s live demonstration and short video enabled the public to gain deeper insight into the impact of their technology. The event also included on-site demonstrations, with research teams directly operating booths so that audience members could experience the innovations firsthand.

In a special public voting segment, an AI Impactor Award was given to the team whose technology and societal impact most impressed audience members. The winner received a KRW 20 million research grant

Follow-up Research Support

 The AI Champion, AutoPhone, will submit a follow-up R&D plan to further advance its technology over 11 months (December 2025–October 2026). Based on the review and approval of this plan, the team will be eligible for up to KRW 3 billion in national R&D funding.

 The AI Innovator, Stratio Korea, will receive up to KRW 1.2 billion for follow-up R&D.

 Each AI Challenger — Revivo, SafeAngel, and FineHost — will be supported with up to KRW 500 million.

Award Name Rank / Number of Teams Prize / Support

Minister of Science and ICT Award

(Champion) 1st place (1 team) Trophy + follow-up R&D funding (up to KRW 3B)

Minister of Science and ICT Award

(Innovator) 2nd place (1 team) Trophy + follow-up R&D funding (up to KRW 1.2B)

Minister of Science and ICT Award

(Challenger) 3rd place (3 teams) Trophy + follow-up R&D funding (up to KRW 0.5B/team)

President of IITP Award 6th–10th place (5 teams) Certificate + KRW 5 million

President of TTA Award 11th–20th place (10 teams) Certificate + KRW 5 million

AI Impactor Award (Popular Vote) 1 team KRW 20 million bonus grant

Winning Teams and Technologies

AI Champion – AutoPhone (Lead: Shin In-sik)

“A mobile AI agent that performs commands by operating various apps like a human user.”

AutoPhone is developing mobile AI agent technology that understands user commands and executes smartphone apps to automate complex tasks, aiming to bridge the digital divide for elderly users and advance the evolution of smartphones into AI-driven devices. As follow-up research, the team plans to further develop “FluidGPT 2.0,” a fully autonomous on-device Large Action Model–based personalized AI agent.

AI Innovator – Stratio Korea (Lead: Lee Je-hyeong)

“Infrared AI: A new era of recycling innovation enabled by custom infrared data”

Stratio Korea aims to overcome the limitations of conventional visible-light–based vision processing through infrared multispectral imaging–based recognition technology, improving material analysis and composite material identification, enhancing efficiency in automated sorting of recyclables, and contributing to the realization of a resource-circulating society. As follow-up research, the team plans to advance its Infrared-AI™ technology by integrating high-resolution infrared imaging systems with AI-based contamination and material analysis algorithms to build a low-cost, high-efficiency automated sorting system.

AI Challenger ① – Revivo (Lead: Yoon Hee-Dong; Daekyo CNS / GURUGEN / WPDR)

“A modular scanning–based intelligent mobility exterior diagnostic system”

Revivo is developing technology that provides an integrated diagnosis of a vehicle’s exterior, underbody, and tire conditions through low-cost sensor–based 3D reconstruction and semantic fusion analysis, AI-based automated precision repair-cost estimation, and modular hardware–based automated vehicle diagnostics with high reliability. As follow-up research, the team plans to establish a virtual repair-cost automation platform by combining modular hardware—integrating laser-line, underbody, and tire scanners—with AI-driven analytical software.

AI Challenger ② / AI Impactor (Popular Vote) – SafeAngel (Lead: Jung Im-doo; UNIST & RIST)

“On-device Physical AI technology for industry-specific active hearing protection and communication devices”

SafeAngel is developing technologies to support worker safety and communication in high-noise industrial environments, including baseline-based anomaly detection, on-device large language model (LLM)-based voice guidance, and active control technologies enabled by Physical AI. As follow-up research, the team plans to complete an AI-based active hearing protection and occupational safety system by combining industry-specific noise databases with customized on-device LLMs and to carry out field demonstrations in actual industrial sites.

AI Challenger ③ – FineHost (Lead: Ji Hyeon-jun; Vacatio)

“An AI solution that automates end-to-end hospitality operations through a voice user interface (VUI)”

FineHost is developing technology that automates the full range of lodging operations through voice commands, including automation of booking, pricing, inventory, and review responses, a LangGraph-based multi-agent structure designed to minimize conflicts, and multilingual autonomous revenue-management technology to ensure reliability in hospitality operations. As follow-up research, the team plans to commercialize a VUI-based hospitality AI platform that integrates multilingual voice-command recognition with autonomous revenue-management algorithms.

Deputy Prime Minister and Minister Bae Kyunghoon stated,

“The AI Champion Competition encourages anyone involved in AI research to take on open innovation challenges. It has become the largest stage for AI innovation in Korea. The Ministry will continue to fully support Korea’s AI talents so that their ambitions can shape the global AI industry in the years ahead.”

For further information, please contact the Public Relations Division (Phone: +82-44-202-4034, E-mail: msitmedia@korea.kr) of the Ministry of Science and ICT. 

Please refer to the attached PDF.