The lawyer belongs at the design table, not the final gate

For twenty years I watched weaponry arrive in military units. Much of it was superb, though some of it carried problems that should have been caught long before anyone fielded it, not because the engineering was poor, but because nobody had asked the legal questions early enough to change the design while changing it was still cheap and straightforward.

That sequencing problem is what I now spend my time fixing.

Most of us would picture (if thinking think about the law of weapons at all) a review at the end: a system is built, tested, ready, and then someone checks whether it is lawful before it goes into service. The obligation to conduct a review, under Article 36 of Additional Protocol I to the Geneva Conventions, sits with States, not with the companies that design and build it. So, it is understandable that a manufacturer would leave the questions regarding lawfulness of use to someone else, as a gate to be cleared at the end of the process.

I’d argue that is the wrong way round.

A weapon has its lawfulness substantially determined by choices that are made at conception: how it discriminates, how its effects are controlled or limited, what happens at the edges of its intended use. By the time a system reaches a formal review, those choices have been fixed in metal and code; supply chain decisions have been made, procurement programmes stretching for years have been established, and marketing and influencing plans are in train. If a problem surfaces then, the options are expensive:

• Redesign;
• restrict its use; or
• absorb the risk.

If the same problem is identified at the design table, it is often a line in a requirements document; it can be solved before it ever costs anything.

So my work runs the length of the lifecycle, not just its end. It starts in the conceptual phase, advising on the legal contours of a design while they can still be shaped. It moves into the design of trials and tests, so that the evidence of compliance is actually generated and captured rather than assumed. It culminates in a structured legal review, one a developer can, if it chooses, share with a procuring State to make that State’s own assessment faster and easier. And for systems that change over time, particularly autonomous and AI-enabled ones, it continues, because a meaningful update can raise the legal questions afresh.

None of this is about slowing development down, or just saying no. It is the opposite. It is a problem-solving exercise that involves identifying effects likely to be judged unlawful, identifying modifications that rectify them and, where that is not possible, removing the offending issue early. This strips risk and rework out of the programme. Designing trials that generate the required evidence and, where it helps, put the right observers in the room means the proof of compliance is ready when a State asks for it, rather than reconstructed under pressure later. Knowing what data will satisfy a procurer, and assembling it deliberately, shortens the path to acceptance. Done well, legal rigour is not a brake on getting to market, it can be the afterburner that propels the system forward.

There is another benefit too, and I will not overstate it. A company that builds the protection of civilians into its engineering, is a company whose staff, shareholders and public can see that it takes that responsibility seriously. In a sector under constant scrutiny, that is not a soft consideration, it is increasingly a commercial one.

I came to this work from two decades as a warfare officer and latterly a lawyer; the combination is deliberate. I have been the end user. I know what it is to be handed a capability and to need confidence in it. And I have the legal knowledge, skills and experience to make sure that confidence is built in from the first drawing, not bolted on at the last.

If you are developing a weapon, a means or a method of warfare and you would rather find the legal questions at the design table than at the final gate, that is precisely the conversation I want to have.

This is the first in a short series for those who design, build and sell military capability: what the law asks of a system, and how answering those questions early strips risk, cost and delay out of a programme. Next: why autonomous and AI-enabled systems are rewriting the timetable.