News

DSEI Connect

02 Aug 2022

Forces Focus RAF - Exploring Human-Machine Teaming

Forces Focus
Forces Focus RAF - Exploring Human-Machine Teaming
Following the Royal Navy (RN)’s first edition of Forces Focus events, which bring the industry and military together to help tackle challenges identified by each Front Line Command (FLC), the Royal Air Force (RAF) hosted the second part of the series. They built on the ideas from RN’s artificial intelligence (AI) and machine learning (ML) discussions to explore Human-Machine Teaming (HMT) opportunities for the RAF.

HMT is the effective integration of humans, robotics and warfighter systems although it extends much further in the military aviation context. It has the potential to transfer the future of battlespace but it comes with a set of problems and roadblocks.

A Woman is speaker during the Forces Focus roundtable discussionsThe participants interacted directly with the RAF and other industry experts during roundtable discussions and a webinar to help solve the challenge posed by Air-Vice Marshal Lincoln Taylor in the Challenge to Industry video. One of the key challenges they identified for maximising the potential and building trust in HMT is the lack of consistency and sufficient collaboration between the industry and the Ministry of Defence (MoD). The industry needs a clear and focused understanding of the requirements, resources, routes and use-cases to maximise the efficiency of their engagement and to enable relevant solutions. There is also a need for universal standards and regulations agreed across all domains that could unlock more frequent opportunities for exploitation that matches MoD’s expectations.

Gaining this clarity is especially difficult for Small-Medium Enterprises (SME) who are facing additional challenges such as: MoD’s contracting mechanisms; Primes locking down the architecture and protecting IP limiting SMEs’ access to end users; feedback and validation of solutions which creates a barrier to collaboration between the three. Hence, it is extremely important to have clear, transparent and focused communication throughout the whole process and between all stakeholders in the supply chain.

 

The same relates to exploiting data, both emerging and already available. Varied algorithms and data analytical techniques could address a wide range of problem sets using just information. Therefore, it is essential to use the existing data, especially in other industries such as data science, agile software, gaming, automotive etc. There are also already unmanned control stations and complex systems being used in admin, back-office and simple procedures. The transformation could be facilitated by drawing on more off-the-shelf technology and applications from other industries to solve defence problems.

 

A speaker standing in front of a whiteboard thinking.Moreover, common standards could unlock more frequent opportunities for exploitation of the data, especially coupled with lowering classification levels. Currently there is a challenge in processing the information quickly, as there is often a delay in data sharing which is even more difficult at the classified level. To speed up that process, it is worth considering encryption although we must weight data timeliness against risks.

However, data is just the beginning. A large-scale physical sandbox is also needed to develop the HMT processes and capabilities in a meaningful way. This would enable running AI or HMT in parallel with existing processes but in a controlled synthetic environment. As a result, it would help to build trust and validate the algorithms.

One of the key challenges in the implementation process is identifying HMT's greatest utility in alleviating the resource burden of routine functions, or in enabling the high-end, complex functions. If the machine focuses on executing routine functions, it may prohibit complex operations but it allows scaling and design flexibility. This is because routine functions will allow humans to focus on completing the complex functions within and outside the team.

On the other hand, having machines perform more complex functions within an HMT may decrease flexibility to react to changing wartime scenarios, whereas routine functions may only require scaling. Therefore, the aim should be for scalable and rapid enhancement to exploit utility in both environments. The transition between the two could involve rapid manufacture of enhanced team-mates, or a process to quickly integrate new algorithms and software.

To achieve that, a team optimisation could be carried out by AI (or another complex function) which implies that an HMT is a self-learning and optimising team. Having transparency in what control is required will be crucial in allowing dynamic intervention across an HMT. This may only be achieved if the developers work closely with the end user and it would be easier in case of utilising routine functions. Additionally, a common open architecture may be also be required to allow the end user to add the human intervention, common interfaces and controls.

 

Tables of representatives discussing Forces Focus topicsThe level of control should be directed by legal and ethical governors to ensure compliance, with the expectation that some circumstances may change during wartime. The debates to establish a coherent framework are already ongoing but there are no clear answers at this stage due to the complexity of the issue. This conversation must progress, with the increase in autonomy to build a level of comfort in using AI-enhanced capabilities. Therefore, the FLC, ethical and legal bodies need to be clear about the ambition, timeline, goals and use cases to shape the narrative as early as possible and agree on international basic concepts.

These propositions suggested during Forces Focus events are one of the first steps into digital transformation and building the HMT trust in defence. They can lead to more efficient use of Cobots which is an autonomous process that lets robots do what they do best and humans to focus on other tasks. Therefore, the RAF is dedicated to working closely with the industry on developing these findings in the aviation context to fulfil their task of developing capabilities.

Intro Text

Following the Royal Navy (RN)’s first edition of Forces Focus events, which bring the industry and military together to help tackle challenges identified by each Front Line Command (FLC), the Royal Air Force (RAF) hosted the second part of the series. They built on the ideas from RN’s artificial intelligence (AI) and machine learning (ML) discussions to explore Human-Machine Teaming (HMT) opportunities for the RAF.
Loading