Kakao and AI Safety Research Institute to Jointly Build AI Safety Evaluation Framework
TECHWORLD ·
✦ AI Summary
Kakao plans to establish a joint evaluation framework with the AI Safety Research Institute to verify the safety of AI models and agents.
The two companies signed a strategic memorandum of understanding (MOU) on the 28th and began collaboration to jointly evaluate AI safety and build an evaluation framework.
The scope of evaluation will expand from language models to multimodal models and AI agents, with Kakao providing infrastructure and the AI Safety Research Institute handling evaluation execution and methodology advancement.
Kakao announced that it will work with the AI Safety Research Institute to establish a joint evaluation framework for verifying the safety of AI models and agents. Kakao and the AI Safety Research Institute signed a strategic memorandum of understanding (MOU) on the 28th. The purpose of the agreement is to jointly evaluate AI safety and build an evaluation framework.
Through this collaboration, the two companies plan to establish an evaluation foundation capable of identifying and verifying potential risks in AI models in advance. Kakao and the AI Safety Research Institute previously carried out a joint safety evaluation of the language model 'Kanana(Kanana)-1.5-9.8B' last year.
Under the agreement, the scope of evaluation will expand from language models to multimodal models and AI agents. Kakao will provide its AI models and agents, and support the infrastructure needed for evaluation and tool development.
The AI Safety Research Institute will be responsible for conducting the evaluations and advancing the methodology. The two companies plan to jointly review the results and discuss the scope of disclosure.
The collaboration will proceed in three phases. In the first phase, safety evaluations will be conducted before and after deployment for language models. In the second phase, the evaluation targets will expand to multimodal language models and AI agents. In the third phase, the companies plan to jointly develop evaluation tools for efficiently verifying the safety of models and agents.
Considering the expanding use of agents, Kakao said it will strengthen its evaluation framework by involving external expert organizations through this collaboration. Based on that, the company said it will continue safety verification throughout the entire process of model development and service implementation. The scope of evaluation will also expand beyond the model itself to risks that may arise during actual task execution.
Kim Se-ung, leader of AI Synergy Performance at Kakao, said the collaboration has made the safety verification framework for Kakao's AI models more objective and rigorous. He also said the company would continue striving to provide AI models and services that take safety and trust into account alongside technological innovation.
Kim Myeong-ju, head of the AI Safety Research Institute, said that as AI technology advances, the importance of a proactive system for identifying and assessing potential risks is growing. Based on that, he said the institute plans to work with Kakao to build an AI safety evaluation framework that meets global standards.
Source: TECHWORLD · Kim Seung-gi
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407499
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.