Who pays when AI fails? Global lawyers debate AI liability and insurance gaps at IPBA

Who should be liable when a self-driving car crashes or an autonomous machine causes harm? That question took centre stage at an Inter Pacific Bar Association (IPBA) panel discussion on robo-liability, where international lawyers examined how law and insurance are struggling to keep pace with artificial intelligence (AI) and autonomous systems..The session was moderated by Kemsley Brennan, Partner at Minter Ellison; and Miriam Rose Ivan Pereira, Counsel at Oh-Ebashi LPC & Partners. The speakers were Sanjay Patel, Barrister at 4 Pump Court; Christian Alexander Mayer, Partner at Noerr; Huzefa Tavawalla, Partner & Head of Digital Disruption at Cyril Amarchand Mangaldas; and Ralph Smith, Counsel at Gómez-Acebo & Pombo. .The discussion began with a fundamental dilemma: traditional legal systems assume a human decision-maker, but autonomous machines challenge that assumption. As the moderators framed it, if a person’s dog bites someone, the owner is liable. But if an AI-driven machine causes damage, the answer is far less clear..Tavawalla explained that India currently has no unified law regulating robotics or AI. He noted that foreign investment in robotics and AI companies is broadly permitted under India’s foreign direct investment regime, with restrictions mainly arising where defence or sensitive geospatial data is involved.On intellectual property, he emphasised that Indian law does not recognise machines as inventors, meaning AI-generated outputs must still be attributed to human creators or operators..From the UK perspective, Patel described how autonomous vehicles are already forcing a rethink of liability principles."Under proposals by the Law Commission of England and Wales, once a user properly engages a vehicle’s autonomous mode, they would not be civilly or criminally liable for accidents occurring while the system is in control. That represents a significant departure from traditional tort law, which typically places responsibility on the driver."Patel predicted that accident disputes will increasingly revolve around sensor data, software behaviour and system logs rather than eyewitness testimony, potentially leading to group litigation against manufacturers where systemic defects are identified..Mayer outlined the European regulatory approach, where autonomous vehicles remain subject to mandatory insurance and strict approval regimes.He explained that EU law now treats software and AI as products under updated product liability rules, expanding potential liability to manufacturers, developers, importers and distributors. Mayer also pointed to a shift in the automotive supply chain, with technology companies gaining influence and traditional manufacturers losing historic dominance..Smith emphasised that despite the uncertainty surrounding AI, existing legal frameworks are not entirely inadequate.He noted that negligence and product liability laws already provide a wide range of potential defendants, including manufacturers, designers, importers and distributors. However, he cautioned that future accidents may become harder for victims to understand."In a self-driving future, an injured passenger may not even know what caused an accident or who to sue," he said. Smith suggested that the growing use of vehicle data and “black box” evidence could fundamentally change how disputes are resolved..He stressed that legal systems are not starting from zero, despite rapid technological change.“We’re all at the beginning of what’s going to be a long journey.”He noted that AI is advancing more rapidly than the legislative process, which explains why people are panicking. At the same time, he said traditional doctrines remain relevant. “There is a very comprehensive legal machinery already in place, just on the old law.Product liability regimes offer a whole panoply of potential defendants.”.A recurring theme was the difficulty insurers face in pricing risk for autonomous systems. Without historical data, insurers may need to create entirely new products combining motor insurance, product liability and technology risk coverage.Some panellists suggested manufacturers may eventually need to maintain ongoing insurance obligations for vehicles even after they are sold..The panel concluded with a warning for the legal profession: as disputes increasingly involve complex technical evidence, lawyers will need to understand the technologies behind autonomous systems in order to advise clients and argue cases effectively.While autonomous systems promise safer transport and workplaces, the legal and insurance frameworks governing them remain in flux. The panel agreed that robo-liability is no longer theoretical; it is already reshaping law, insurance and legal practice worldwide..Follow Bar and Bench channel on WhatsAppDownload the Bar and Bench Mobile app for instant access to the latest legal news.