MRS publishes AI guidance

The report, called The BEST Framework for Gen AI: How and when to build AI into your research processes, provides a practical framework to help researchers make strategic decisions on when and how to use AI in their projects and where to target investment.

The framework aims to classify the different types of problems researchers seek to tackle into four quadrants – boost, expand, shift and transform – and identify the best type of AI to use to help achieve the project’s objectives.

For example, the most common form, boost, focuses on automating and optimising well-established and predictable processes, while transform requires building understanding from scratch, making sense of unpredictable environments and relying on emergent data and insights.

Alongside the report, MRS is publishing its new ethics guidance on AI, which aims to help researchers consider the ethical design, use and outputs of AI to mitigate harmful biases introduced by the input data on which the technology relies.

The guidance will be regularly updated to reflect the rapidly evolving policy and regulatory frameworks in this area, and formal training will also soon be offered by MRS.

Jane Frost, chief executive at MRS, said: “There are a lot of promises and predictions around AI completely transforming everything we do overnight. In reality, much of what is possible with AI at the moment is a continuation of the work that research and insight have been doing for some time with machine learning and other tech.

“There are of course risks involved with AI, which we have to be alive to. I’m confident, however, that with the right tools and guidance, the sector – which is adept at handling personal data and adopting new technologies – will be able to bring together the power of machines and the creativity and curiosity of human minds for the significant enhancement of our insights.”