The field of research automation is moving towards increased collaboration between human and artificial intelligence agents. This shift is driven by the need for more efficient and transparent research processes. Recent developments have focused on creating multi-agent frameworks that can automate tasks such as literature curation, idea generation, and peer review. These frameworks have shown promise in generating diverse and evidence-aligned hypotheses, as well as improving the accuracy of test oracles in software engineering. Noteworthy papers in this area include:

• Autoresearcher, which introduces a multi-agent demo system for knowledge-grounded and transparent ideation.
• HIKMA, which presents a framework for semi-autonomous scientific conferences that integrates AI into the academic publishing and presentation pipeline.
• Nexus, which generates test oracles through a structured process of deliberation, validation, and iterative self-refinement.