Talks and presentations.

The talk read the Swedish Supreme Court’s decision in Anton Foley and others v Sweden (the Aurora case), Ö 7177-23, 19 February 2025, against the ITLOS Advisory Opinion of 21 May 2024 and the ICJ Advisory Opinion of 23 July 2025, and asked what obligations the resulting framework imposes on states in respect of negative emissions. Sub-seabed CO₂ storage furnished the empirical test. Sweden was the analytically useful subject because it stands in two capacities at once: as the respondent in its first systemic climate case, in which the Supreme Court declared the claim inadmissible on grounds of standing and the separation of powers; and as a state that, having concluded cross-border arrangements under the provisionally applied 2009 amendment to the London Protocol in April 2024, now authorises the export of CO₂ for storage beneath the Norwegian continental shelf.

The talk advanced two connected claims. First, that the pollution-transformation principle articulated by ITLOS at paragraph 231, read with Article 195 UNCLOS, supplied a doctrinal basis for treating the long-term integrity of a storage site as part of the exporting state’s stringent due-diligence obligation under Article 194, rather than as a matter exhausted at the permitting stage; and that this reading, while it does not condemn authorised and well-performing storage, locates with some precision the point at which the obligation bites. Second, that the liability architecture on which negative emissions now depend is structurally immature at exactly that point: the transfer-of-responsibility mechanism in Article 18 of the EU CCS Directive has never been completed at commercial scale, and the international regime contains no equivalent mechanism at all.

The talk concluded that this doubled gap, rather than the outcome of the domestic litigation, is the strongest form of the systemic argument the Aurora claimants advance, and that it exposes a mismatch between rights-based domestic adjudication, which reasons from individual harm, and the inter-State and erga omnes logic of the advisory opinions.

Following a single tonne of CO₂ from a corporate pledge through capture, cross-border transport, sub-seabed storage and the long tail of post-closure liability, the lecture argued that a net-zero claim is only as credible as the law standing behind each handoff, and that, at present, none of those layers is finished. The opening anchor was a stocktake of the world’s largest listed companies: roughly seventy per cent of Forbes Global 2000 revenue now sits under a net-zero target, but only about seven per cent of those targets meet basic integrity criteria. The remainder rest on capture and storage as the implicit residual.

Tracking the tonne, the lecture set out four cracks in the enabling law. First, three overlapping disclosure regimes (ISSB, EU ETS, and India’s BRSR) measure the same physical tonne against meters that can be wrong for years, as at Sleipner, the world’s oldest storage site, where a faulty flow transmitter saw injected volumes over-reported by roughly a quarter before correction. Second, the 2009 amendment to Article 6 of the London Protocol, which authorises cross-border export of CO₂ for sub-seabed storage, remains, seventeen years on, not in force; European shipping proceeds under a provisional resolution and a handful of bilateral arrangements. Third, the pollution-mitigation paradox left open at paragraph 231 of the ITLOS Climate Advisory Opinion (Case No. 31, 2024), which warns against transforming one type of pollution into another, was framed by the Tribunal in respect of marine geoengineering rather than CCS; reading it across to sub-seabed storage is an extension by analogy the lecture pressed, not the Tribunal’s own holding. Fourth, the transfer-of-responsibility procedure in Article 18 of the EU CCS Directive, the most developed long-tail liability bridge in any jurisdiction, has never once been used, because no commercial storage site has yet closed.

Beneath all four sits the physical floor identified by Gidden et al. (Nature, 2025): a prudent global geological storage budget of roughly 1,460 Gt CO₂ once seismic, hydrological, protected and populated areas are excluded, an order of magnitude smaller than older technical estimates, and capable, used optimally, of contributing only about 0.7 °C of cooling. The closing argument framed the lecture’s diagnostic as four portable questions to put to any net-zero pledge anywhere in the world: who counts the carbon, who governs its passage across borders, who holds it when the company that made the promise is gone, and is there any room left in the vault at all.

Carbon dioxide removal is no longer a speculative proposition. The IPCC (2022) identified it as a necessary component of pathways that limit warming to 1.5°C, and among the various CDR approaches under consideration, those operating in the marine environment have attracted particular scientific and commercial attention in recent years. Yet virtually all of this activity has taken place within national waters; the legal framework that would govern marine CDR (mCDR) in areas beyond national jurisdiction remained fragmented and internally contradictory.

Two recent legal developments have reshaped the regulatory landscape for mCDR in ABNJ, while a third, longer-running process has failed to keep pace. The BBNJ Agreement entered into force on 17 January 2026, establishing for the first time a binding environmental impact assessment framework for the high seas and the Area. Seven months before that, ITLOS delivered its Advisory Opinion in Case No. 31, characterising anthropogenic greenhouse gas emissions as pollution of the marine environment under UNCLOS and imposing on States Parties a due diligence obligation it described as ‘stringent.’ Against these advances, the London Protocol regime, administered by the IMO rather than under UNCLOS, continues its slow and incomplete evolution.

The talk argued that resolving the central tension, namely that mCDR may both serve states’ due diligence obligations to reduce marine pollution from greenhouse gases and itself constitute a new form of marine pollution under UNCLOS Article 1(1)(4), required treating the BBNJ Agreement’s EIA provisions as the procedural backbone and the ITLOS opinion’s due diligence standard as the normative compass, while drawing on the London Protocol’s technical assessment frameworks for substantive guidance on individual techniques. The Saya de Malha Bank in the South West Indian Ocean served as the case study.

The closing panel of the CCS at Sea conference drew together the day’s threads on the regulatory and governance challenges of carbon capture and storage at sea: cross-border transport of CO₂, multi-source pipeline cargoes and stream composition rules, the assignment of long-term stewardship and leakage liability, and the boundary between EU CCS Directive obligations and the international regimes (London Protocol, OSPAR, BBNJ) that overlap with it. The discussion took the structural question forward: how the law should distribute the legal risk of an offshore CO₂ infrastructure that no single instrument was designed to govern.

Drawing on Rudolf von Jhering’s purposive jurisprudence, the talk argued that the regulation of CCS at sea has been hampered by an inherited tendency to reason from category rather than from purpose. A single molecule of captured CO₂, on a single voyage from a Norwegian capture plant to a sub-seabed reservoir, acquires at least eight legal identities along the way: emission, captured CO₂, cargo, potentially noxious substance, waste, dumping, storage certificate, avoided marine pollution. Each identity belongs to a different regulatory regime that was designed without the others in mind.

Through analysis of the London Protocol, the BBNJ Agreement, and the May 2024 ITLOS Advisory Opinion, the talk considered how the legal architecture might better serve its proper end, climate mitigation, by treating captured CO₂ as a function-defined activity rather than as a substance to be forced into pre-existing categories of waste, cargo, or pollutant. The Jheringian critique of Begriffsjurisprudenz (conceptual jurisprudence) supplied the diagnostic vocabulary; the talk closed by sketching what a purpose-first reading of the existing instruments would look like in practice.

The talk examined an emerging challenge in the CCS value chain: how to allocate responsibility when CO₂ from multiple emitters is physically blended and transported as a single maritime cargo. As Northern European CCS networks scale toward multi-source hubs and cross-border shipping routes, the physical commingling of CO₂ streams creates attribution problems across three regulatory domains.

First, the EU ETS establishes installation-level attribution through monitoring, reporting, and verification requirements, allowing operators to deduct captured CO₂ from compliance obligations. However, when blended cargoes suffer transport losses, the framework provides no mechanism for proportionately allocating leaked tonnage back to contributing emitters for allowance surrender. If five emitters ship blended CO₂ and two percent leaks in transit, existing rules cannot determine who surrenders allowances for those lost tonnes.

Second, multi-source operations expose private law gaps in contractual risk allocation. When individual CO₂ streams meet purity specifications but the blended cargo is rejected as off-spec, no established liability framework determines responsibility between contributing emitters, terminal operators, or transporters. Questions of contamination liability, indemnification for storage site damage, and force majeure allocation remain unresolved where fault cannot be attributed to a single party.

Third, cross-border shipments create jurisdictional ambiguity under instruments like the London Protocol and OSPAR. When cargo originates from multiple states, determining the ‘exporting State,’ allocating responsibility through the transport chain, and resolving whether ETS deductions apply for storage outside the ETS jurisdiction remain unclear.

The talk concluded that while existing frameworks support single-source CCS, multi-source operations require targeted regulatory clarification or standardised contractual frameworks to preserve accountability where physical traceability has been lost through blending.

The closing session of the colloquium took up the question that runs from animal-cruelty statutes to the design of social robots: should the criminal law sanction acts that cause no concrete harm to a person but still violate a widely-held moral intuition? The talk situated this within the twentieth-century Hart–Devlin debate on legal moralism (should the criminal law punish behaviour the community finds morally wrong but which produces no provable harm?), and then asked whether anthropomorphism reopens the question for a new class of non-human entities.

Anthropomorphism, on the talk’s account, is not a recent phenomenon emerging with humanoid robots; it is one of the oldest human survival mechanisms, often reflexive and bypassing conscious evaluation. Under current Turkish criminal law, the destruction of an anthropomorphic robot is treated as nothing more than damage to property, a categorisation that does not match the moral reaction the same act provokes from observers. Drawing the through-line from animal cruelty laws (which protect creatures whose subjective suffering we cannot fully verify) to humanoid AI agents and social robots, the talk closed by sketching the contours of a Turkish criminal-law response that takes anthropomorphism seriously as a category-disrupting fact, without collapsing into either pure moralism or a thin property-protection paradigm.

The talk situated surrogacy at the intersection of medicine, ethics, and law. Framed against International Women’s Day, it opened on the autonomy questions that surrogacy raises for the women who carry, then traced the history of assisted reproduction from the pharmaceutical production of estrogen in 1936 through the first IVF birth in 1978, the first gestational surrogacy in 1985–86, the Baby M and Calvert v. Johnson cases, and Ukraine’s 2009 surrogacy law. From there it turned to the regulatory question for Turkey, addressing surrogacy as a hope for families (including queer families) for whom the existing Turkish legal framework has no place, set against the commercialisation and exploitation risks that the practice carries when women’s bodies become the medium of family formation.

An important area of application of robotics technologies is unmanned water-surface and underwater vehicles, for example in remote exploration work. The talk focused on the consequences of these technologies in a liability context: that because of the characteristics of these vehicles, especially autonomy, their development and increased use in the civil sector are likely to require a new approach other than the well-established fault-based liability regime, but that they are not expected to require amendments to the basic tenets of maritime law as illustrated, for instance, in the 1972 IMO COLREGs Convention. The argument was that even though most unmanned water-surface and underwater vehicles are becoming more and more autonomous, they are closer to reasonable safety when compared to the ultra-hazardous activity of unmanned aerial vehicles. The argument was developed further in the 2021 Springer chapter Lessons Learned from Robotics and AI in a Liability Context.

The conference paper that became the chapter Anthropomorphic Machines (Banja Luka 2020). The talk argued that human-robot interactions differ structurally from interactions with other artefacts because robots are autonomous, and that recent advances have enabled robots to take roles formerly thought of as human-only: companions, lovers, social interaction partners. Robot anthropomorphism, the attribution of human attitudes and emotions to robots, generates implications for individuals and society in the longer term, from the manipulation of emotional attachments to an expansion of existing privacy risks. The talk called for legal orders to stop treating robots as mere tools and to start treating them as parties to interactions, laying the groundwork for examining the legal problems arising from robot autonomy.

The talk treated robots as artefacts that work according to the sense-think-act paradigm, and argued that with their growing autonomy and emergence as alternatives to human companionship (pets, friends, confidants, romantic partners) they confront the self-understanding of human beings at both the individual and collective level, and the legal system that is designed in line with that self-understanding. From there it worked through the legal questions this opens up, chiefly anthropomorphism as a structural feature of human-machine interaction rather than an error to regulate away, and the implications for the legal order. The argument provided early ground for the doctoral work that followed.