Skip to content

University of Bristol's hand in the future of AI

AI’s energy use is a powerful limiting factor on development, writes Miles Gilroy.

By Miles Gilroy, First year, Astrophysics

A recent wave of investments from UK Research and Innovation (UKRI) into the development of AI has seen the University of Bristol receive £21 million in funding to set up two research hubs, with the aim of developing AI in areas that benefit the general population and propelling the UK into a dominant position in the rapidly evolving field of artificial intelligence.

The Engineering and Physical Sciences Research Council (EPSRC), part of UKRI, has invested a total of £80 million in nine of these research hubs in universities across the UK, including the universities of Bristol, Edinburgh, Cardiff, and many more. A further ten scoping studies have been funded by the Arts and Humanities Research Council (AHRC) to explore the ethical and responsible integration of AI across areas including education, policing, and the creative industries.

Of the nine hubs funded by the EPSRC, three will focus on the mathematical and computational groundwork required for the development of AI, whilst the remaining 6 will explore the integration of AI into scientific industries.

The University of Bristol has taken the lead on two of these hubs: AI for Collective Intelligence (AI4CI) and Information Theory for Distributed AI (INFORMED-AI). 

The AI4CI hub aims to take real-time data and use it to develop machine learning and smart agent technologies to benefit individuals and national agencies in areas such as healthcare, pandemics, cities, finance, and the environment. The plan is to use AI to enhance collective and individual decision making, and streamline the process of defining tailored solutions and policies to issues like global pandemics or treatment plans for individuals suffering from common conditions.

The INFORMED-AI hub is targeted at developing the foundations that will allow multiple AI systems to work together, focusing on the efficacy, resilience, and reliability of their collaboration.

The AI4CI hub is part of the six hubs focused on the integration of AI into industry whilst the INFORMED-AI hub is part of the six focused on the foundations of AI systems. The benefit of running these two hubs simultaneously is that when INFORMED-AI has conducted sufficient research and development into the creation of AI, AI4CI has already performed the groundwork for introducing AI into the world in an effective yet responsible way. This means the process is extremely efficient, ensuring the UK remains at the forefront of the artificial intelligence industry.

Despite the profound benefits AI could have on society, it does come with a multitude of issues that these research hubs aim to keep firmly in mind as they develop it.

The security of AI is a subject of heavy debate. Is it wise to depend so heavily on technology, especially with our economy and healthcare systems, in a world where our personal data and livelihoods are already threatened by cybercrime? An idea that UKRI is keen to promote is the development of AI that can be trusted and relied upon to deal with such pivotal foundations of society. This not only means AI that is resilient in the face of malicious attacks, but also AI that improves upon the reliability of humans in the same position.

However, this leads to the concern of human redundancy. If AI can perform specific tasks better than its human equivalents as well as being able to effortlessly collaborate and interact with other versions of itself, then what is the point of having humans in the roles that AI will come to occupy? Obviously the ultimate aim of the research hubs, such as those at the University of Bristol, is to lay out the foundations for the development of AI that can enhance human productivity and can integrate into the inner mechanics of society to improve human existence. Finding a way to do this without creating AI that replaces humanity is of utmost importance and is arguably the main aim of this research.

There is also the impact that the use of AI could have on the environment if left unchecked. It is no secret that the processing power required to run AI systems is immense, and with that comes an immense demand for electricity. Despite current strides towards greener energy, the majority of electricity generated still comes from fossil fuels. Not only does this mean that AI will contribute massively towards climate change and the contamination of the atmosphere and oceans, but, according to Sam Altman, CEO of OpenAI, an energy breakthrough is necessary for the existence of AI, which will require vastly more power than people have expected.

This presents a powerful limiting factor on AI development that needs to be overcome before we can even imagine an AI driven future. Innovations in the fields of renewable energies are all well and good, but the energy yield of source such as solar or hydroelectric are far too small for the future we are heading towards, and governments are still reluctant to rely on nuclear fission due to its harmful waste products and catastrophic (yet unlikely) historical incidents.

This leaves one option: nuclear fusion. Nuclear fusion is the process that produces energy within the sun. It is by far the most efficient and cleanest source of energy possible, but unfortunately we are unable to harness its power on Earth yet. This means investing in nuclear fusion is arguably more important for the development of AI than investing in the development itself, at the moment. However, this does not mean the creation of the research hubs has come too early as having the knowledge already available as soon as we can use it streamlines its integration into society. 

In the case that AI development is ‘complete’ before fusion development, it is always possible to limit its use by prioritising industrial or governmental applications and then release its potential to everyone else once we have the facilities to maintain it. 

These are all concerns that the research hubs funded by UKRI aim to address in their surveys of the development and integration of AI.

Featured image: Unsplash / Emile Perron

Do you think these new developments by the University of Bristol will change the job market you graduate into in the short term?