Artificial Intelligence: Vaccine Technology

Question

Asked by

To ask His Majesty’s Government what assessment they have made of the potential for artificial intelligence to contribute to developments in vaccine technology.

The results from Cambridge and the DIOSynVax programme, primarily funded by the Government through Innovate UK, is a good example of how the Government are taking AI’s potential in vaccine technology seriously. AI’s most credible contribution to vaccine development is in the design stage, including predicting antigen structure and binding and predicting how pathogens evade immunity. Our £137 million AI for science strategy is funding some key enablers for AI use in drug discovery, including in medical research and engineering biology.

My Lords, I thank my noble friend for that Answer. If the House does not realise, what has happened in Cambridge is the first use of AI to design a vaccine candidate being used in human trials, with the possibility of being able to deal with a broad range of viruses. One of the researchers involved said that it is

“amazing what we can do with it for the good of humanity”.

Unfortunately, this new technology may also, in turn, potentially pose a great threat to humanity. So I ask my noble friend: what work is going on to ensure that this type of development is subject to the strictest controls and that, in future, no bad actors can in any way use this new technology to pose a threat to humanity?

I completely agree that this technology is hugely important for the good it can do in making vaccines and new medicines. It will change our ability to treat and prevent diseases. It is also true that the same technology can be used for bad things. Like many technologies, it can be misused. Many things are in place already: it is illegal, of course, to make chemical and biological weapons, and the Government take very seriously the threat that can come from new technologies. The biological security strategy was published last year. There is £580 million in the defence investment plan for Porton Down infrastructure for a national biosecurity centre, and £115 million to strengthen defences against AI risk. This is a very active area of work.

My Lords, the development of Evo 2—which is an AI biological model with 9.3 trillion nucleotides extending over 128,000 different species, including humans—means that there is now a capability to produce proteins, molecular structures and even genomes, and therefore artificial pathogens. So is it time for the UK to have mandatory enforcement and regulation of functional-base synthetic DNA?

I thank the noble Lord for his question. The point about the technology is that an enormous amount of good comes from it, and we must not forget that. But yes, there are risks, and we are looking at nucleic acid synthesis using the programmes now available to do that, and we are looking at what can be regulated. Initially it will be a voluntary process, but we are looking at the options to understand and control the use of new synthesis of materials through nucleic acids.

My Lords, we should all recognise the immense potential of AI in this field. But, on regulation, will the Minister explicitly confirm and ensure that a human-in-command approach will apply to any use of AI in vaccine approvals and pharmacovigilance, so that decisions on approvals and safety monitoring are always taken by qualified human experts, rather than delegated to algorithms or an automated system alone?

The MHRA is one of the most important and best regulators in the world for medicines, and none of the requirements for an AI-assisted vaccine design is different from those for a normal one. In other words, it has to be looked at in the same way, with the same sort of safety requirements. The MHRA is looking at how it can use AI, but certainly as part of a human-linked process. There will be people making these decisions.

My Lords, the best recent example of vaccine technology is the pandemic and Covid, and the UK rightly champions the 100-day mission to develop vaccines against any future pandemics. As I am sure the Minister is aware, AI can rapidly compress that discovery phase through genomic surveillance, lab validation and manufacturing. Given that, how is AI being incorporated into the 100-day mission?

As the noble Lord says, the 100-day mission was endorsed at the 2021 G7, and I was very involved in that process. AI will be a critical part of how we do vaccines, therapeutics and diagnostics, and I am optimistic that this makes the 100-day mission really achievable, particularly for vaccines. But—and this is an important but—not for all organisms: it is not always the case that you can get a vaccine.

My Lords, if this technology is used for malign purposes, it does not really matter whether that happens in the UK or elsewhere in the world. What action is being taken and how confident are we that there is scope for international agreement, particularly among the major players on the international scene, to develop the necessary safeguards to protect all of us on this planet?

This is a very hot topic of discussion. A lot is going on between nations and it is a big topic at the G7, where the UK is leading on some of these discussions.

My Lords, we know that public confidence in vaccines is essential if there is to be good uptake and if they are to be effective. What are the Government planning to do to ensure that there is public confidence in AI-generated vaccines, and have they considered involving community and faith groups in this? Those are key to ensuring public confidence in this area.

Public acceptance of vaccination is very high in this country, and that is built on trust, part of which comes from the regulator. The regulator in this country is extremely rigorous and will make sure that all the usual things that a vaccine needs to go through will be gone through, whether or not AI has been involved in the process somewhere. There will be no change in the regulations as a result of that, except where there may be opportunities to get better safety profiles because of the use of AI. That is important. As noted, it is crucial that public engagement with this happens. The MHRA is good at doing that, and it will be an important part of continuing to build vaccine confidence in the UK.

My Lords, I welcome the potential for AI in developing vaccine technology. However, AI models are as good as the data we put in to train them. I therefore have two questions for my noble friend the Minister. What steps are the Government taking to close the data gap for those from ethnic minorities and for women and pregnant women to be included in trials? How are they working with international partners, including the WHO, to expand the database so that we achieve health equity for all?