The Group of Governmental Experts on lethal autonomous weapons systems (GGE on LAWS) has adopted 'lifecycle measures' — legal reviews, risk consideration, testing and bias mitigation — in paragraphs 36 to 39 of its final report covering the 2024–2026 CCW sessions. The measures apply at the study, design, development, acquisition and adoption stages of an autonomous weapon, reaching earlier in a system's life than the point of use, which anchors the preceding paragraphs of the adopted elements.

Netta Goussac, Associate Senior Researcher at the Stockholm International Peace Research Institute and Associate Partner at Lexbridge, analysed the elements in a symposium post. A GGE participant since 2017 and a member of the Australian delegation during the second session of 2026, she writes that the measures 'facilitate a state's adherence to the prohibitions and constraints on use of LAWS' and 'convert a legal obligation or policy commitment into practical and tangible actions by states.'

Where does the 'lifecycle' concept come from?

The lifecycle concept 'dawned' in the debate in 2018 and appeared in the GGE's 2019 Guiding Principles. Principle (b) states that human responsibility for decisions on the use of autonomous weapons should be considered across the lifecycle, while principle (g) specifies that risk assessments and mitigation measures should be part of the 'design, development, testing and deployment cycle of emerging technologies in any weapons systems.' Australia contributed a nine-stage 'system of control' to the Group in 2019, running from the legal and policy framework to after-action evaluation.

The drafting trail shows steady evolution. The July 2024 rolling text carried a short list headed 'other measures to ensure compliance with [international humanitarian law]'. By November 2024 this had become a discrete block, introduced by a chapeau covering measures taken 'prior to potential employment and as applicable throughout the entire life cycle of LAWS.' By June 2026 the block split: two obligations framed with 'must', the remainder qualified by 'should… as appropriate, across the life cycle of LAWS.'

The adopted elements contemplate four kinds of lifecycle measures:

• Legal reviews (paragraph 36)
• Consideration of risks (paragraph 37)
• Measures ensuring capabilities and limitations can be understood (paragraph 38)
• Measures against bias (paragraph 39)

A fifth measure, in paragraph 42, addresses human-machine interfaces, guidance and training in the accountability cluster.

What does paragraph 36 require?

The first sentence of paragraph 36 reads: 'In accordance with States' obligations under applicable international law, in the study, development, acquisition, or adoption of a new LAWS, determination must be made whether its employment would, in some or all circumstances, be prohibited by international law.'

The sentence mirrors guiding principle (e) of 2019, substituting 'a new LAWS' for 'a new weapon, means or method of warfare' and inserting 'applicable'. The verb oscillated between 'must' and 'should' throughout the rolling text before the GGE returned to the 2019 formula. That consistency reflects legal diversity: not every State Party to the Convention on Certain Conventional Weapons is party to Additional Protocol I of the Geneva Conventions, whose article 36 contains the legal review obligation. Some non-parties — the United States and Israel most prominently — review capabilities as a matter of domestic policy.

Paragraph 36 is still innovative. No CCW protocol requires or recommends legal review as an implementation measure, and neither does the Anti-Personnel Mine Ban Convention nor the Convention on Cluster Munitions. Any instrument incorporating the element would be the first weapon-specific instrument to restate this general rule of international humanitarian law. Its practical force may be limited: few states maintain a standing review mechanism, and LAWS are demanding to review.

What happened on modification reviews?

The adopted text says nothing about reviewing modified systems — a consequential deletion. Sub-paragraph 39(b) of the Chair's suggestions would have required review of any further development or modification with a 'substantial impact on the operational functioning' of the identification, selection and engagement functions, or on anticipated effects. Sub-paragraph 38(c) would have prevented mission parameters being substantially modified without human authorisation. Neither survived adoption. Some states preferred 'material impact' over 'substantial'; the United States sought deletion or its own formulation, confined to modifications that could render use prohibited or substantially change the circumstances of lawful use — despite US Department of Defense Directive 3000.09 and US Army Regulation 27-53 both requiring senior or legal review upon substantial changes domestically.

What did survive is encouragement of voluntary exchange of relevant practices, subject to national security and proprietary-information considerations — the only genuinely new part of paragraph 36 relative to 2019.

How are risks and testing addressed?

Paragraph 37 requires that risks of civilian casualties and precautions minimising incidental loss of civilian life, injury and damage 'must be considered' during design, development, testing and deployment. Other risks — unintended engagements, loss of control, proliferation, acquisition by terrorist groups — fall under 'as appropriate'. The internal split separates IHL compliance risks, framed as mandatory, from security risks, framed as discretionary.

Paragraph 38(a) now frames testing and evaluation, 'including adequately realistic simulations', as necessary 'to assess how the LAWS will perform in the expected operational environment.' The singular 'expected operational environment' replaced the Chair's 'variable operational environments', narrowing testing back toward anticipated circumstances of use. Paragraph 38(b) addresses 'guidance, training and instructions' for 'relevant personnel, including those who bear responsibility in the chain of human command and control' — the broadest formulation yet. Requirements that LAWS operate predictably and reliably, with explainable and traceable functioning, fell away; delegations divided on whether these were legal or policy terms, and the text resolves the argument by omission.

What changed on bias?

Paragraph 39 asks states to prevent and mitigate 'unintended' bias — in data sets, artificial intelligence models and other technological components — and, separately, automation bias in the use of LAWS. The qualifier changed from 'harmful' to 'unintended' in May 2025. The Chair's summary records that most delegations welcomed the change as more neutral, while some observed that a degree of bias is intrinsic to distinguishing targets. The vocabulary is imported from AI governance and appears in the US-led Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy and the Pathways to Action document of the 2026 REAIM Summit.

Why do these measures matter for practitioners?

Paragraph 42 locates interfaces, guidance and training in the accountability cluster, and the location is instructive. Before an attack, these measures enable compliance with distinction, proportionality and precautions. After one, they make it possible to establish what happened — the foundation for investigating and prosecuting war crimes under the grave breaches regime or the Rome Statute's complementarity arrangements. As Goussac puts it: 'What states do before use determines what they are able to know afterwards.'

The lifecycle measures are the standards by which the effectiveness of any instrument on LAWS will be judged. Since the GGE commenced work in 2017, developing states have elaborated policies, guidelines and standards, and in some cases publicised their interpretation of IHL obligations. That prompting of national practice, Goussac writes, 'is one of its durable achievements.'