The new wave of generative AI tools and capabilities is challenging many longstanding traditions, beliefs and ways of working in universities and research labs. Professionally speaking, for many of us it is likely to herald the biggest disruption we encounter in our lifetimes, and possibly the biggest disruption societally as well. The ‘quantitative revolution’ will look tame by comparison, in terms of both scope and speed. All academic fields of inquiry will likely be impacted, not just those that are considered data rich. This talk will explore the capabilities of some of the more experimental and research-focussed AI methods currently in use, including: (i) agents that can determine if a research claim is supported by the literature, (ii) science large language models that have been specifically to solve specific research challenges, (iii) AI tools to automatically document research data with the right subject-level metadata, (iv) AI agents that can write research articles and (v) AI systems that can make original scientific discoveries. Progress here challenges many aspects of our traditional approaches to science, even the need for models based on theory, since models based solely on data can now provide more accurate and faster results. As AI progress accelerates, the implications for research and for researchers may become even more profound. The implications for universities, and for the pursuit of knowledge will be considered and will hopefully lead to a fruitful discussion with those attending.