Generative AI in Software Engineering: Ongoing research and emerging issues

The proliferation of disruptive large language models in recent years has enabled a plethora of applications across several domains. In software engineering, LLMs have shown remarkable capability in understanding and generating software artifacts. In our research group we have applied LLMs to multiple SE tasks such as detecting malicious code and summarising software documentation. Despite their promising performance, LLMs still face several challenges: hallucination remains a major concern; as LLMs increasingly rely on external sources such as retrieval-augmented generation knowledge bases, they become more susceptible to adversarial attacks disguised in fine-tuning data; trained on vast datasets including public code repositories and security databases, LLMs can effectively reproduce patterns from their pretraining data — introducing a critical risk of data memorisation. The agenda of the seminar: (i) brief introduction to our current research on the applications of generative AI in software engineering; (ii) discussion of ongoing research topics, including our recent work on probing memorisation in LLMs.