A research and innovation laboratory at the university will attempt to develop safe artificial intelligence (AI) solutions and systems.
The initiative is expected to allow the university to “push the forefront of generative AI” as well as “benefit society and stimulate economic growth” by powering applications in areas including robotics, drug discovery, medical diagnoses, semiconductor development and tackling climate change.
GAIL’s researchers will also work in partnership with the university’s Centre for Technomoral Futures to address the ethical, legal and regulatory frameworks necessary to ensure the safe and responsible use of AI, the university said.
“GAIL builds on world-class research at the School of Informatics in AI and machine learning, which has been applied to a variety of fields such as healthcare, fintech and sustainability,” said Professor Helen Hastie, head of the university’s School of Informatics.
“The school is well placed to be an anchor for the GAIL cross-disciplinary research and development by forging key collaborations within the university and with external partners and policymakers.”
“Edinburgh is uniquely placed to provide world-class leadership and expertise around generative AI,” said Professor Iain Gordon, head of the College of Science and Engineering at the university. “Once operational, this exciting initiative will create a new pipeline of talent with the skills to both benefit from and shape the economic and societal benefits of this transformational technology.”
The announcement comes amid a dramatic rise in the popularity of foundation models, a category of AI systems trained on huge volumes of data such as text, images, video or audio to gain broad and sophisticated capabilities across many tasks.
The capabilities of these next-generation tools have allowed researchers to create powerful AI tools such as OpenAI’s ChatGPT and Midjourney. However, governments and experts have raised concerns about the risks these tools could pose to people’s privacy, human rights or safety.
The news follows the announcement by the UK government that Edinburgh will host a trio of AI training centres from UK Research and Innovation, including one in biomedical innovation, one in language processing and one in robotics.
Last month, Prime Minister Rishi Sunak revealed the UK will launch an AI safety institute, following a similar announcement from the US.