{"id":285714,"date":"2026-09-09T17:24:09","date_gmt":"2026-09-09T17:24:09","guid":{"rendered":"https:\/\/www.europesays.com\/iran\/285714\/"},"modified":"2026-09-09T17:24:09","modified_gmt":"2026-09-09T17:24:09","slug":"compliance-by-design-v-loac-in-the-israeli-defense-industry","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/iran\/285714\/","title":{"rendered":"Compliance by Design V: LOAC in the Israeli Defense Industry"},"content":{"rendered":"

Editors\u2019 note: This is the tenth post in a series dedicated to Lethal Autonomous Weapons Systems (LAWS) and the questions of human oversight and legal accountability under international humanitarian law. Previous posts have focused on LAWS<\/a>,\u00a0China<\/a>,\u00a0Russia<\/a>, the\u00a0United States<\/a>, and\u00a0Europe<\/a>. The second part of the series examines defense industry practices. It analyzes how hardware developers, software providers, and manufacturers approach the integration of law of armed conflict requirements into autonomous combat systems in the\u00a0United States (see here<\/a> and here<\/a>), China<\/a>, Europe<\/a>, and Israel.<\/p>\n

The last post in a series is the natural point to take stock of where the argument has landed, if it landed at all. If nothing else, it certainly covered a fair amount of airspace. At just over 18,000 words, the series has grown to roughly the length of a short Victorian novella, like The Strange Case of Dr Jekyll and Mr Hyde whose exploration<\/a> of duality resonates well with the subject<\/a> of these posts.<\/p>\n

But while autonomous systems are an elusive subject, they are neither inherently virtuous nor inherently malign. The same drone can reduce civilian harm through greater precision<\/a> or can also magnify destruction if poorly designed, inadequately tested, or employed in ways that disregard<\/a> legal and operational safeguards. The challenge explored throughout the second half of this series has been how to ensure that autonomous weapons remain closer to Jekyll than Hyde, not by assumption, but through Compliance by Design.<\/p>\n

The Israeli Defense Tech Powerhouse<\/p>\n

As the third entry<\/a> focused on Moscow\u2019s stance on Meaningful Human Control (MHC), the question as to why the last post in the series is dedicated to Israel instead of Russia conjures up naturally. The notion that it makes sense to put the Lethal Autonomous Weapons Systems (LAWS) industry of a country of just over ten million people under the microscope instead of a great power<\/a> might appear striking at first glance. It, however, is not. Although understanding Russia\u2019s position within the Convention on Certain Conventional Weapons (CCW) Group of Governmental Experts<\/a> is important to grasp the debate on autonomous weapons, publicly available information on the design, testing, and verification of Russian autonomous systems is sparse, making it difficult to assess whether, and to what extent, law of armed conflict (LOAC) considerations are incorporated through Compliance by Design.<\/p>\n

By contrast, from technical descriptions to doctrinal statements, Israeli companies and the Israel Defense Forces (IDF) produce a relatively substantial body of public materials. To understand why, one has to go back to the 1950s. This was the decade when Israel started to invest heavily in defense R&D, while highly specialized signals\u2011intelligence units such as Unit 8200<\/a> began training generations of engineers and spawned dual-use technologies that later spun out into civilian firms. During the 1990s, the dissolution of the Soviet Union brought the arrival of more than 800,000 immigrants<\/a> to the country and with them the doubling of Israel\u2019s technical talent pool and a turbocharging of the country\u2019s technology sector<\/a>. The USSR had 99 problems, but STEM education<\/a> was not one.<\/p>\n

Israel absorbed this influx through two programs: a 1991 incubator<\/a> for immigrant researchers; and the 1993 Yozma<\/a> which used public funds to attract foreign, especially American, venture capital and Silicon Valley expertise. Coupled with defense industry downsizing releasing engineers into the labor market, these policies led to the birth of not only one of the world\u2019s most important high-tech clusters, the Silicon Wadi<\/a>, but also transformed Israel into a defense\u2011tech powerhouse<\/a>.<\/p>\n

LOAC, Article 36, and Israel\u2019s practice<\/p>\n

The sector that made the country into one of the world\u2019s<\/a> leading arms<\/a> and indeed drone exporters<\/a> is building on two important features: tightly integrated relationship<\/a> between manufacturers; and the IDF<\/a> and fielding autonomous and semi\u2011autonomous systems under sustained operational demand.<\/p>\n

Israel is a High Contracting Party to the 1949 Geneva Conventions<\/a> and has long maintained that the law of armed conflict governs the development and use of all weapons, including those incorporating autonomous functions. The country\u2019s official position<\/a> affirms that international humanitarian law<\/a> applies fully to systems with autonomous capabilities and that human judgment and responsibility remain essential across the life cycle of any weapon.<\/p>\n

However, Israel is not a party to Additional Protocol I<\/a> and therefore is not bound by the formal Article 36<\/a> legal review obligation. Nevertheless, the country has stated<\/a> that legal review<\/a> of new weapons constitutes the best available instrument for ensuring that only lawful means of warfare are employed, and it conducts such reviews as a matter of domestic practice that are functionally similar to Article 36 reviews.<\/p>\n

Given this framework the key question is how Israel\u2019s industry practice translates LOAC into design. Does the tight IDF\u2013industry integration and sustained operational demand lead manufacturers to embed LOAC considerations of distinction<\/a>,\u00a0proportionality<\/a>,\u00a0precaution<\/a>, and\u00a0humanity<\/a> early, technically, and verifiably, ensuring that these principles become concrete technical requirements? Or do operator oversight and post\u2011design procedures, refined through operational experience, shape legal compliance heavily?<\/p>\n

Industry Rhetoric vs Design Reality<\/p>\n

The majority of Israel\u2019s autonomous and semi-autonomous platforms are manufactured by the three firms that dominate the country\u2019s defense industry: Israel Aerospace Industries (IAI); Rafael Advanced Defense Systems; and Elbit Systems. While the latter is publicly traded, IAI and Rafael remain wholly State-owned although both plan<\/a> to go public. These companies work in a highly distinctive<\/a> close institutional proximity to the Ministry of Defense and the IDF, and their development cycles are frequently shaped by immediate operational requirements<\/a>.<\/p>\n

Israeli firms present systems with autonomous functions in their product descriptions<\/a> as operating under human-in-the-loop or man-in-the-loop architectures, with the operator retaining authority over the decision to engage and the capacity to abort. Yet, as with the State\u2019s position, the precise quality and extent of human involvement required is left open.<\/p>\n

Three Platforms, One Design Logic<\/p>\n

Considering Israel\u2019s capacity in advanced technology and defense innovation, it would be reasonable to expect developmental programs analogous to the U.S. collaborative combat aircraft efforts<\/a> and associated autonomy software architectures<\/a>. Yet there is no such comparable, publicly-documented Israeli program. While this might strike some as quite odd, the reasons are comparatively simple. Continuous<\/a> operational demand<\/a> has directed Israeli industry effort toward systems that can be refined under combat conditions rather than toward long-lead, next-generation air combat platforms.<\/p>\n

Rafael\u2019s L\u2011Spike family<\/a> which has been exported to 41 countries and expanded in 2025 with the L\u2011Spike 1x<\/a> and L\u2011Spike 4X<\/a>, integrates AI\u2011enabled functions for automatic target recognition and target image acquisition during the loitering<\/a> phase. The architecture is described as man\u2011in\u2011the\u2011loop: the operator receives dual\u2011mode (Electro-Optical\/Infrared<\/a>) seeker video; conducts identification; and authorizes the terminal dive, with hardened data links intended to preserve this control link under contested and GPS\u2011denied conditions<\/a>. An operator can supervise and coordinate up to four missiles simultaneously, while retaining decision authority over engagement. Although these features support operator judgment at the point of engagement, they do not constitute independent technical constraints that enforce distinction or proportionality at the design stage.<\/p>\n

Israel Aerospace Industries\u2019 Harop<\/a> is a standoff loitering munition supervised by remote human\u2011in\u2011the\u2011loop mission control. It conducts autonomous search and loitering within a designated area, with the attack being able to be aborted if required. Related Harpy\u2011family systems, including the Mini\u2011Harpy<\/a>, offer selectable modes that include higher autonomy against radiating targets and man\u2011in\u2011the\u2011loop engagement. The abort capability and operator supervision provide mechanisms for intervention, but publicly available materials do not indicate the incorporation of fixed, auditable discrimination or proportionality thresholds into the system architecture itself.<\/p>\n

Elbit Systems\u2019 SkyStriker<\/a> employs autonomous navigation during cruise and loiter phases together with an automatic video tracker. Engagement is directed against operator\u2011designated targets, with an abort function that permits the munition to go around and re\u2011engage. While the system is described as \u201cfully autonomous\u201d in navigation and tracking, the decision to engage remains with the operator. The design prioritizes operator authority over the decision to use force rather than embedding LOAC principles as independent technical features from the outset. The principal safeguards across these platforms remain real\u2011time sensor feeds, operator supervision, and abort authority.<\/p>\n

Operational Performance, Export Footprint, and LOAC<\/p>\n

The list above leads directly to the obvious question of operational history. Unlike developmental LAWS programs whose compliance features remain largely<\/a> prospective, Israeli platforms possess combat records both in Israel and in other countries<\/a> and a significant export footprint. Their record makes it possible to move well beyond design intentions and assess how a model centered on retained human judgment, real-time supervision, and abort authority performs in practice. In this architecture, autonomy is applied strictly to navigation, persistence, and detection; the decision to apply lethal force remains entirely a human one.<\/p>\n

The overall picture is, however, quite uneven. Operational records confirm tactical effectiveness<\/a> against military targets under a range of conditions. Yet they expose the contingency of an architecture that places the decisive legal judgments\u2014distinction, proportionality, and the decision to abort\u2014almost entirely, real time, on the operator. When data links remain intact and engagement rates remain manageable, the design can support discriminate use. Under conditions of electronic warfare, compressed decision cycles, high operational tempo<\/a>, in other words circumstances that have a strong tendency to materialize during high intensity combat<\/a> and may also be coupled with dense urban terrain<\/a>, the same design offers few independent technical backstops. The capacity to intervene or abort is only as reliable as the operator\u2019s situational awareness and the integrity of the control link at the critical moment.<\/p>\n

Measured against the standards of Compliance by Design this approach preserves a direct channel for human legal judgment but does not translate the core principles of LOAC into early, fixed, and auditable constraints<\/a> on system behavior. As a limitation, this becomes more salient when systems are exported to operators with different training and oversight regimes, that is, virtually every time a system is exported. Indeed, it would be unrealistic to expect design assumptions, training standards, and institutional oversight calibrated to one force do not automatically accompany the platform. Absent fixed discrimination thresholds, auditable proportionality metrics, or other design stage constraints independent of the operator, lawful performance depends heavily on factors external<\/a> to the system itself.<\/p>\n

As a case study, Israel illustrates a fairly distinct industrial pathway. It is one defined by continuous, combat-driven refinement of autonomous functions paired with retained human authority, rather than the formal embedding of legal constraints at the earliest stages of design. Compliance remains an operator practice, not a technical property of the system.<\/p>\n

Conclusion: Compliance by Design as a Spectrum<\/p>\n

The series has examined LAWS from two complementary perspectives: first, how the United States, Europe, China, and Russia interpret human responsibility and accountability under the Law of Armed Conflict; and second, whether those legal commitments are translated into the design, testing, and fielding of autonomous weapons through Compliance by Design.<\/p>\n

States diverge not only in their understanding of Meaningful Human Control and the necessity of additional regulation but also the extent to which legal obligations are transformed into concrete engineering requirements. Compliance by Design is therefore far from emerging as a neat, uniform practice, but as a spectrum ranging from early integration of legal safeguards into system architecture to approaches that continue to rely primarily on operator judgment at the point of engagement.<\/p>\n

The United States demonstrates the most mature publicly documented examples of embedding legal review into software architectures, testing regimes, and system development from the outset. Europe has sought to institutionalize similar principles through Article 36 reviews, procurement requirements, and funding conditions, although fragmentation and regulatory complexity continue to present significant obstacles. China has prioritized rapid capability development within its opaque civil-military ecosystem where verifiable design stage safeguards remain difficult to assess. Lastly, Israel provides yet another model, one shaped by continuous operational demand, rapid combat-driven adaptation, and close integration between industry and the military. Compliance depends predominantly on retained human judgment rather than on fixed architectural constraints.<\/p>\n

Taken together, the findings of the series strongly suggest that the central challenge facing autonomous weapons is not whether humans remain legally responsible. Among major militaries and in public CCW statements, there is remarkably little disagreement on that point. The more consequential question is whether legal obligations are embedded early enough, deeply enough, and transparently enough to influence how autonomous systems actually function under combat pressure. Human responsibility provides the legal foundation. Compliance by Design determines how robustly that responsibility can be exercised in practice.<\/p>\n

Ultimately, autonomous weapons are neither inherently lawful nor inherently unlawful. Their compliance with LOAC will depend less on abstract declarations of human responsibility than on the engineering choices made long before a system reaches the battlefield. As autonomy is bound to rapidly expand across military operations, legal accountability will depend far less on last-second human decisions and more on whether distinction, proportionality, precaution, and meaningful human control have already been engineered into the weapon itself. The future of lawful autonomy will be decided not on battlefields, but at the drawing board.<\/p>\n

***<\/p>\n

Dr Gerald Mako<\/a> is a Research Affiliate at the Cambridge Central Asia Forum at Cambridge University.<\/p>\n

The views expressed are those of the author, and do not necessarily reflect the official position of the United States Military Academy, Department of the Army, or Department of Defense.<\/p>\n

Articles of War\u00a0is a forum for professionals to share opinions and cultivate ideas.\u00a0Articles of War\u00a0does not screen articles to fit a\u00a0particular editorial\u00a0agenda, nor endorse or advocate material that is published. Authorship does not indicate affiliation with\u00a0Articles of War, the Lieber Institute, or the United States Military Academy West Point.<\/p>\n

\u00a0<\/p>\n

\u00a0<\/p>\n

\u00a0<\/p>\n

\u00a0<\/p>\n

\u00a0<\/p>\n

\u00a0<\/p>\n

Photo credit: Julian Herzog via Wikimedia Commons<\/p>\n

\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"Editors\u2019 note: This is the tenth post in a series dedicated to Lethal Autonomous Weapons Systems (LAWS) and…\n","protected":false},"author":2,"featured_media":285715,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[33],"tags":[10279,45185,24175,77500,2382,711,6327,77501,37,48551,77502,77503],"class_list":["post-285714","post","type-post","status-publish","format-standard","has-post-thumbnail","category-israel","tag-ai","tag-ap-i","tag-compliance","tag-compliance-by-design","tag-defense-industry","tag-drone","tag-geneva-conventions","tag-gerald-mako","tag-israel","tag-laws","tag-lethal-autonomous-weapons-system","tag-loac"],"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@iran\/117242250230627603","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/posts\/285714","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/comments?post=285714"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/posts\/285714\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/media\/285715"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/media?parent=285714"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/categories?post=285714"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/iran\/wp-json\/wp\/v2\/tags?post=285714"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}