KakaoBank said Tuesday that four of its research papers on financial artificial intelligence (AI) were accepted to international journals this year, highlighting the internet-only lender's efforts to improve the safety and reliability of AI use in finance. According to the company, the papers were accepted to the International Conference on Learning Representations (ICLR), the Association for Computational Linguistics (ACL) and the Language Resources and Evaluation Conference (LREC). The research focuses on improving the security, accuracy and safety of generative AI models for financial services. One paper, presented at ICLR in April, introduced a system for detecting prompt injection attacks targeting large language models used in specialized fields such as finance and law. Developed jointly with The Korea Advanced Institute of Science and Technology (KAIST), the model was trained on a proprietary dataset of about 59,000 samples and outperformed existing models in detecting malicious prompts. Two more papers were presented at LREC in May. One proposed a method for detecting prompt injec
Kakao Bank AI safety research gains global recognition