Role of Community-Based Computing Services
The UK Government's recent investment of £900 million in the UK AI Research Resource (AIRR) aims to provide world-class compute resources to UK-based researchers, promoting AI activities that are more likely to be safe, sustainable, and socially beneficial. However, the challenges in implementing public compute policies for AI development are substantial.
The UK possesses only 1.4% of total global supercomputer capacity, ranking 10th in the world. The largest AI models today use 10 billion times more training compute than the largest models in 2010, and the amount of training compute used in the largest models is doubling every five to six months. This rapid growth in compute requirements highlights the need for increased public investment in AI infrastructure.
One of the main challenges in implementing public compute policies is ensuring equitable access to computational resources. Without public policies, AI development tends to be dominated by Big Tech firms that control the large-scale compute resources necessary for training AI models. This concentration impedes an open and competitive AI ecosystem.
Another key challenge is preventing monopolization and private enclosure of AI infrastructure. Providing public access to powerful compute resources raises concerns about data security, misuse, and maintaining strong data governance for accuracy and trustworthiness of AI outcomes.
Governments can pursue several strategies to overcome these challenges. Building and maintaining public AI compute infrastructure is crucial to fostering a diverse and competitive AI R&D environment. Promoting open-source and collaborative AI research helps democratize access to AI tools and prevent enclosure by proprietary interests.
Establishing robust data governance frameworks is also essential to improve data quality, interoperability, and security standards to ensure reliable AI outputs and maintain public trust. Developing clear, adaptive regulatory sandboxes can create domain-specific regulatory environments where AI innovations can be safely tested under flexible rules to accelerate adoption while ensuring accountability.
Implementing AI explainability and transparency standards is necessary to make sure AI decisions in public applications are understandable to stakeholders to uphold fairness and trust. Investing in workforce development is crucial to address AI talent shortages through training and hiring programs to support the design, deployment, and oversight of public AI infrastructure.
The goal is to create an AI ecosystem that is open, innovative, trustworthy, and aligned with public interests rather than dominated by narrow commercial or ideological motives. This requires coordinated investment, regulatory foresight, and commitment to transparency and inclusivity in AI compute policy by governments.
However, it's important to note that a clear vision of AI that links its functioning in the UK economy to a wider vision about the society we want to live in is currently lacking. The governance of AIRR should center the perspectives of users and relevant stakeholders, including researchers, small and medium-sized enterprises, and frontline professionals.
In the short term, AIRR could use commercial cloud services to rapidly meet existing demand, leveraging the buying power of the UK Government. Public capital expenditure is currently circumscribed by tight fiscal, infrastructural, and ecological limits. Therefore, public compute investments should be framed as an industrial policy lever for reshaping the dynamics of AI development and promoting the creation of public value throughout the AI supply chain.
The proposed timescale of public compute investments has been called into question due to extensive wait times for advanced hardware and the public sector's relative lack of agility in procurement. Public compute policies should be coupled with a wider suite of industrial policy measures to help steer the AI market, including pro-competitive measures, investments in monitoring infrastructure, and new institutions at the data layer.
The UK Government could set longer-term targets for onshoring the different stages of the compute supply chain to build diverse domestic (including public) capacity. This would help ensure the UK's AI sector remains competitive and innovative in the face of growing global competition.
National industrial strategies are being focused on the AI sector in various jurisdictions. Notably, France is investing significant sums in public compute, and the establishment of a National AI Research Resource (NAIR) is a long-term bipartisan policy goal in the USA. The success of these strategies will depend on their ability to balance the need for innovation with the need for transparency, inclusivity, and accountability.
References:
[1] Kak, A., & West, S. M. (2021). National industrial strategies for AI: A critique from the left. The Journal of Industrial Policy, 20(2), 107-128.
[2] Myers West, S., & Kak, A. (2021). The AI industrial policy toolkit: A left-wing approach to AI development. Technology in Society, 69, 102277.
[3] The Royal Society (2017). AI in the UK: Ready, Willing and Able?. London: The Royal Society.
[4] The Royal Society (2021). AI in the UK: The Road to 2030. London: The Royal Society.
[5] The Alan Turing Institute (2020). AI Strategy for the UK. London: The Alan Turing Institute.
- The UK Government's recent investment in the UK AI Research Resource (AIRR) underscores the relevance of data-and-cloud-computing technology in finance and business, as it aims to empower UK-based researchers with world-class resources, fostering safe, sustainable, and socially beneficial AI activities.
- To ensure that the benefits of AI development in the industry are equitable and inclusive, it's essential to implement public compute policies that promote open-source and collaborative AI research, investing in workforce development, and maintaining robust data governance frameworks, all while mitigating concerns about data security and misuse.
