Written evidence submitted by Scale AI.
Scale AI is pleased to respond to the Defence Sub-Committee on Developing AI Capability and Expertise in UK Defence’s call for evidence. We look forward to working with the Sub-committee, and the entirety of the UK Government to ensure that the UK continues to be a global leader in the adoption of artificial intelligence.
Introduction
Scale was founded in 2016, with the mission to accelerate the development of AI applications by helping organisations integrate AI and machine learning into their operations and workflows. From our earliest days working with the leading autonomous vehicle companies to our work today with the leading tech companies like OpenAI, Microsoft and Meta, Scale has always been at the forefront of AI innovation. Today, we fine-tune the data for, red team, or test and evaluate nearly every Large Language Model before it comes to market.
Beyond our work in the commercial sector, we also have experience and knowledge of the defence sector. We have worked closely with the U.S. government since 2019 to provide artificial intelligence and machine-learning solutions and translate deep data sets into insights for a plethora of defence-related functions. These include intelligence, surveillance, and reconnaissance services and analysis for critical administrative and operational U.S. government and military platforms.
Our experience working on cutting edge innovations for both the commercial and defence sectors uniquely positions us to lend our expertise to this call for evidence.
1) How clearly has the Ministry of Defence set out its priorities for the kind of AI capacity and expertise it believes the UK defence sector should have, what priorities has it identified, and are these deliverable?
Scale AI strongly supports the UK government’s recognition that AI will play a key role in homeland defence and conflicts abroad and we welcome the government’s emphasis on artificial intelligence within its defence policy agenda. Due to this level of interest and support, Scale AI has been engaging with the UK Ministry of Defence (MoD) to explore how we can support the UK’s defence objectives in order to ensure that the UK defence sector is not just keeping up with technological advances in the field, but rather leading the way forward on how artificial intelligence and machine learning can be integrated into and significantly improve defence functions across the board.
The MoD Defence Artificial Intelligence Strategy, published in June 2022, rightly identifies that Artificial Intelligence is one of the technologies essential to defence modernisation. The four objectives identified in that strategy set the right direction by recognising the need to strengthen the broader UK ecosystem, which will enable AI adoption, and acknowledging the need for the UK to collaborate with international allies.
The third objective aims to address commercial barriers and incentivise engagement and co-creation. This is the right approach, as for the government’s objectives to succeed, the expertise of leading private companies is essential. Scale AI’s initial engagement with UK officials has been positive, with officials across the defence and security functions of the civil service willing to engage and share insights into how to foster tech adoption across government.
However, despite positive engagement with a range of officials, the UK procurement process remains somewhat opaque. It is imperative that there are clear paths to enable companies, like Scale, to move beyond briefings and demonstrations into pilot programs and fielding. If this path does not exist, the UK risks seeing leading defence technology companies potentially depriotise the UK market instead choosing to work with more forward leaning governments.
The extent to which the government can successfully embed AI into the MoD at pace hinges on adopting a whole-of-system approach where MoD teams across sub-departments collaborate between functions. In the short and medium term, prioritising the enablers of digital transformation and adopting easily implementable AI technologies (or ‘AI Now’) will ensure later projects are not held up.
This challenge is not unique, as the U.S DoD has had similar challenges, but governments who most quickly adapt to them will reap the benefits that accompany leading the world in the adoption of AI.
2) What strengths and expertise does UK industry currently have in the field of Artificial Intelligence with defence applications?
Scale AI welcomes the UK’s wholehearted embrace of the potential of AI, thanks largely in part to the commendable efforts of the UK Government in prioritising this agenda. As a company with unique insights into AI scalability, we are actively sharing our knowledge with key government departments, including Number 10 and the Department for Science, Innovation & Technology, to promote the UK as a global AI hub.
The UK’s long standing history in cutting edge AI research, paired with your pro-business environment for companies, like Scale AI, uniquely position the UK to be a leader in AI adoption. This leadership was recently on display with the hosting of the global summit on AI in the UK, which Scale AI was pleased to attend. Over the course of the two days, this Summitt helped to bring together international allies and industry leaders to initiate discussions on how AI can be developed safely. The Bletchley Declaration was the first step for international collaboration on frontier AI safety, including how best to support national and international frameworks.
The National AI Strategy published in December 2022 recognises that building on the UK’s strengths in AI will take a whole-of-society effort that will span the next decade. This will involve investment in the long-term needs of the AI ecosystem and approach which seeks to attract and train the AI skills that the workforce will need to ensure a whole system approach. The government’s investment of £54 million in AI and data science workforce and development is a welcome sign that the UK is committed to establishing the wider ecosystem which is necessary to fully harness the transformational opportunity of AI.
The creation of the AI Safety Institute, stemming from the AI Safety Summit, underscores the significance of Research & Development (R&D). This in conjunction with a strong academic and research community, particularly in defence applications, has played a pivotal role in establishing the UK’s global edge. The UK’s outward-looking approach, collaborating on the development of frameworks and capabilities in partnership with allies, ensures its agility and progression in an ever-evolving ecosystem.
The UK has adopted a strategy towards AI, extending its focus to enhance defence capabilities. It’s also critical that this strategy turns into action and implementation. The rest of the world is not slowing down, and the countries that most efficiently embrace AI will lead the world in the adoption of it. While the UK, to date, has been in a leading role, this leadership is by no means guaranteed in the future.
3) How can the UK Government best develop capacity and expertise within domestic industry in sectors such as engineering and software to support the development and delivery of Artificial Intelligence applications in defence?
To realise its ambition, the government must attract the most exceptional minds in the AI sector. Seeking to create this expertise in isolation would be challenging, so it is welcome that the government has sought to embed international cooperation at the heart of its approach to AI development. For more than seven years, Scale has worked on the forefront of nearly every AI innovation, and we stand ready and willing to partner with the UK government to support its progress towards AI adoption.
The UK should also seek to establish home-grown talent to ensure a thriving AI ecosystem is created in the UK. Scale AI runs ‘Centers of Excellence’, physical hubs that run end-to-end ML lifecycle (technical experts, data labellers, ML engineers, SMEs, infrastructure support staff) through our state-of-the-art technology. These hubs are fantastic areas to build an AI workforce from the ground up and democratise the technology as many of the positions require only a high school degree. We are looking to expand our Centers of Excellence programme into the UK to help foster domestic expertise. The MoD has signed several MOUs with industry to facilitate digital skills training for the defence sector. As part of a wider programme Scale AI would be interested in exploring this as a potential avenue to initiate our work with the UK government.
Beyond the workforce, the UK government must also ensure that they are best set up to harness the power of AI. AI is only as good as the data that it is trained on, and every day the UK government generates mass quantities of data, much of it goes wasted. However, in order to lead the world in the adoption of AI, steps must be taken to convert this raw data into “AI-ready” data. Countries around the world have embraced this concept, which the US Government calls “AI Scaffolding” which sees raw data tagged, annotated and labelled in order to strengthen AI systems.
In addition to creating higher quality AI systems, this will also help to ensure a joint-up approach to development and avoid silos between departments. Alongside this, the government’s focus should go beyond the raw technology aspect of AI development and seek to foster all the other areas within the ecosystem, including setting a balanced approach to regulation, a clear plan for workforce development and an ethics framework. This includes the importance of open data sets for AI research and expertise and investment in the AI workforce through lifelong education and training.
4) What can the Government do to help embed UK AI companies in defence supply chains, both domestically and internationally?
Embedding UK AI companies in defence supply chains, both domestically and internationally, requires a strategic cross-government approach. Collaboration between the United Kingdom, the United States, and other allied nations in the realm of artificial intelligence is not only advantageous but essential. Scale AI encourages the development of interoperable frameworks to enable effective AI cooperation, as well as facilitating the realisation of the benefits of AI.
Such interoperability does not require complete uniformity of domestic approaches, given that there may be a need for targeted approaches based on national circumstances and applicable legal frameworks. Interoperable approaches should be supported by appropriate standardisation and, where suitable, shared principles, codes, or similar frameworks.
The AI conference hosted by the UK was a welcome initial step for countries to agree on a shared set of principles and a plan to promote interoperability and knowledge sharing between nations. This builds on work from international actors like the G7 and the OECD, for like-minded democracies of the world will commit to common AI principles reflecting our shared values and priorities.
By maximising this cooperation, nations will be better at unleashing AI innovation. Importantly, Scale AI would be recommended that governments take a holistic approach to AI, with a strong emphasis on international, cross-cutting collaboration among governments, the private sector, and academia throughout. To do this, the UK government should focus on flexible, light-touch policy environments that support AI research and development and encourage innovation and competition.
5) How can the UK Government ensure that it champions the UK AI sector in the context of Pillar 2 of the AUKUS Partnership?
As mentioned previously, the cooperation of like-minded countries is vital for the UK and its allies to gain a strategic advantage in military settings. To this end, Pillar 2 of the AUKUS agreement is a welcome development, with a focus on sharing technology and increasing interoperability between their armed forces. Fostering deeper integration of security and defence-related science, technology, industrial bases, and supply chains will enable this cooperation to be a success.
A clear plan of which platforms and capabilities are intended to be discussed and studied within the AUKUS framework is not yet fully substantiated, but it remains a sensible vehicle for developing platforms and capabilities that all three allies wish to share. A robust AUKUS framework should be designed to allow innovation in artificial intelligence, quantum, and other key defence technologies to flourish.
To enable this project to succeed, AUKUS needs clear support from leadership at senior levels of all three respective governments. Further objectives and clear metrics should be developed to enable the progress of the working groups will be important to ensuring the program’s success. Within this framework, there needs to be a clear consideration of how the AUKUS partners intend to integrate the shared knowledge and expertise of private companies such as Scale AI.
CONCLUSION
Scale AI is grateful for the opportunity to respond to the Call for Evidence, and we look forward to continuing to work with the UK government to position the UK, and the MoD, as a global leader in the adoption of artificial intelligence.
30th January 2024