Peer questions
- What makes internal versus external data sharing uniquely difficult in aerospace and defence?
- How can organisations mitigate the tension between data sovereignty and the need to collaborate?
- What cultural and behavioural blockers persist, even where legal and technical solutions exist?
- What models exist for trusted, decentralised data sharing across supply chains?
- Can industry-level collaboration accelerate standardisation and interoperability?
Making internal data sharing the default
One aerospace firm applies a policy called “open by consent”: internal data is presumed shareable unless explicitly restricted. Open by consent works on the notion that data is presumed shareable, but only with explicit and informed consent from the data subject (unless there are legal or ethical reasons to override this).
This flips the burden of justification and prevents siloed hoarding. “For internal data, you need to justify why you cannot share it,” said a senior data governance expert. The leader found that this approach maintained compliance and documentation of everything while avoiding sharing blockages. That shift accelerated uptake, especially when backed by executive sponsorship.
Culture, not compliance, is the main blocker
While legal constraints like the International Traffic in Arms Regulations (ITAR) – a US government legislation controlling trade in defence articles – and export control are real and must be adhered to, many leaders agreed the bigger challenge is behavioural and cultural. “The safe answer is always ‘no’,” said one attendee. Reluctance to make a change or evolve a system is often rooted in habit or fear, not regulation. Even when data can be shared, users default to legacy platforms or avoid perceived risks.
Bridge the knowledge gap to change behaviour
Unlike sectors where commercial motives impede data sharing, defence organisations face a different cultural issue: people want to do the right thing, but they are not sure how. “I think the quandary is trying to do the right thing, but not really knowing what the right thing is,” said one participant. To combat this, data and AI leaders need to secure investment in education and clarity around classification and consent to convert agreement into action.
Jurisdictions cause sharing complexities, even internally
Even intra-group sharing can become complex when organisations span jurisdictions. A CTO described how inside the UK, sharing is seen as relatively smooth. But when sharing involves crossing borders, data and AI leaders are effectively sharing externally, even within their own company. This complicates policy and language alignment. There must be a high level understand of different national regulations from numerous departments and leaders in a multinational corporation which requires further investment and time considerations.
Ensure semantic clarity for correct shared data use
Some data producers resist sharing due to fear of misinterpretation. “Sometimes you cannot understand my data, and I would prefer you don’t have it, because you will make bad decisions because you don’t understand it,” said one contributor. Semantic clarity and access become critical to avoid incorrect usage and blame.
Symmetry is essential for successful cross-functional sharing
Structured data often flows from design to engineering to support, but value is realised much later. “Contributing data where you see the benefit is one thing, but sharing data for the benefit of others in the organisation is something else. When input, process, and output in an organisation are asymmetric, it is far more difficult to achieve alignment around sharing,” observed one attendee. The asymmetry of effort and reward needs strategic resolution.
Decentralisation can cause governance problems
Pooling sensitive data centrally is not feasible in the defence sector and comes with serious concerns regarding governance and ownership. One participant explained their pivot towards a decentralised model: “We federated services in the middle that will help organise data, including trust frameworks and managing data sharing agreements covering the data provider and the data consumer’s access and usage policies. But data being decentralised and having one architecture to pull all the data causes a governance issue: who is the one governing and honing the architecture in the middle?”
Standardisation is the sector’s Achilles’ heel
While all attendees agreed on the need for digital continuity and structured data exchange, the industry lacks common semantics and technical standards. “A key challenge is that our sector is not good with standardisation. We like things made by ourselves and consistently over engineering stuff, making it difficult to align on one simple way to achieve success as an industry,” one attendee remarked.
Without alignment, scalable collaboration is impossible. It is essential that data and AI leaders in the defence industry come together regularly to discuss and develop digital structured data exchanges.
Cultural change to improve data sharing
Attendees discussed how technologies enabling column and row-based permissions exist, but uptake is patchy. As one data and AI leader put it, “When someone says, ‘can I access that data? Can I share it? Can I put a copy in the lake house?’ The safe option is to say no, and that’s the barrier we’ve got to try and get over. If you’re allowed to access it in source, you should be able to access it from the lake house. But the culture says ‘well, let me keep hold of it’.” The industry is in a situation where capability is ahead of maturity and a cultural shift – which notoriously takes time – is required to rectify this bottleneck.
Look to other sectors for models and pitfalls
The group welcomed the idea of learning from open banking’s trust frameworks and market architectures. “There’s nothing like that in our industry. It is more linear in supply chain terms, and the controls on data sharing,” said one participant. Crucially, any model must scale from bilateral agreements to decentralised ecosystems without compromising control.
What data and AI leaders should do
- Formalise share by default policies backed by executive enforcement.
- Train teams and leaders on semantic understandings in addition to legal compliance.
- Create cross-functional data usage maps to surface asymmetries and bottlenecks.
- Co-develop a decentralised data sharing architecture with partners.
- Define lightweight trust frameworks and enforceable access controls.
- Engage industry peers in standards-setting, especially for structured data.
- Pilot internal marketplaces for low-risk data to build confidence.
- Audit and streamline existing bilateral agreements to reduce redundancy.
- Avoid over-reliance on central platforms and focus on interoperable, federated services.
- Explore follow-up sessions with other sector pioneers to understand scalable governance models.
