Second Workshop on the Responsible Use of Generative AI in Research

Second Workshop on the Responsible Use of Generative AI in Research

This workshop is the second in a series of three workshops designed to understand UCD staff and researchers use of GenAI technology and appreciate their level of understanding regarding the ethical use of GenAI tools. This research project is funded by UCD’s ReCLAIM funding scheme. The first workshop was held in the School of Computer Science in September 2024 and the third workshop will be organised in Spring 2025.

On the afternoon of Wednesday, 13th November, CeADAR staff gathered in Seminar Room G.21 at CeADAR offices within the NexusUCD building located within the Belfield Office Park for an interactive workshop titled “Responsible Use of Generative AI for Daily Research Activities”. The event was held on the same day as CeADAR’s town hall meeting to ensure more staff had the opportunity to participate in this in-person event to discuss the evolving role of AI in industry-oriented research settings.

Photos to go here

(opens in a new window)https://photos.google.com/share/AF1QipMTU8GvzhK33QlEbmhOaHkLGvvulBqivfkDOjS4L59SYoIPahQl-CGB1bjbIv85Hw?pli=1&key=Z2ZRNkZPMXVLeUgzanhIUkRMMmRIeVdiWmwyS0VR

(Courtesy of Conor Cosgrove)

The workshop began with a presentation by Assistant Professor Shen Wang, the School of Computer Science’s Research Integrity Champion, who outlined its key objectives. These included promoting the “Living Guidelines on the Responsible Use of Generative AI in Research” (https://research-and-innovation.ec.europa.eu/document/download/2b6cf7e5-36ac-41cb-aab5-0d32050143dc_en?filename=ec_rtd_ai-guidelines.pdf) and gathering concrete feedback and examples from CeADAR data scientists, operations team and some invited guests from CeADAR partner organisations. The insights collected are intended to help researchers across UCD enhance their productivity while upholding the highest standards of research integrity.

Key Findings from the Pre-Workshop Survey

A summary of survey results from 12 responders — primarily data scientists and principal investigators (>50%) — offered interesting insights into current GenAI usage. For example:

• ChatGPT (both free and paid versions) emerged as the most widely used tool, favoured by two thirds of responders, surpassing tools like Gemini, Claude and Copilot, which was expected to dominate coding-related tasks.
• More than 80% of respondents agreed that GenAI positively impacts their research with roughly 8% expressing doubts or hesitations due to concerns about research integrity. These numbers are in relatively stark contrast to the School of Computer Science respondents were just over 50% agreed that GenAI can positively impact their research and more than 30% expressed reservations.
• Popular uses of GenAI include text rewriting/summarisation/generation whereas code generation/debugging, advanced information retrieval, and image generation saw less adoption.
• Two thirds of responders had not undergone any formal training on using GenAI tools and this compared to 80% of Computer Science respondents.

Discussions and Insights

Following the survey overview, Dr. Adrian Byrne (CeADAR) and Dr. Shen Wang facilitated group discussions centred on the guidelines for the responsible use of GenAI. Researchers found some recommendations, such as “Refrain from using GenAI tools for sensitive activities e.g. peer-reviews or evaluations”, “Maintain a critical approach to using GenAI and continuously learn how to use it responsibly to gain and maintain AI literacy”, and “Remain ultimately responsible for scientific output” to be relatively straightforward to adopt.

The discussions also clarified more nuanced guidelines, such as: “Use GenAI preserving privacy, confidentiality, and intellectual property rights on both inputs and outputs” and “When using generative AI, respect applicable national, EU and international legislation, as in their regular research activities.”.

Open-ended remarks sparked lively debate, including:

• “More detailed guidance on how to transparently report the usage of GenAI tools in research would be helpful.”
• “The reporting guidance available at the moment is quite vague.”
• “Let it rip! But it's very context dependant.”

Future Workshops

This workshop marked the second event under the ReCLAIM Round 1 project (https://www.ucd.ie/cs/news/ucdcomputerscienceresearchersreceivereclaimawardforgenaiproject/). The next and final session will take place in Spring 2025 and will focus on the responsible use of GenAI across all UCD schools (excluding Computer Science).