Revised Product Liability Directive (Directive (EU) 2024/2853)

Operative mechanics: strict liability re-engineered for software and AI

The revised Directive (EU) 2024/2853 keeps the 1985 regime's no-fault core but redefines its perimeter for the digital economy. Art. 4(1) brings "software" — including standalone AI systems — within the meaning of "product," and Recital 13 treats an AI system provider within the meaning of Regulation (EU) 2024/1689 as a manufacturer irrespective of delivery model. The redress machinery is the load-bearing element: Art. 6 fixes compensable damage (death, personal injury including medically recognised psychological harm, property, and destruction or corruption of non-professional data); Art. 8 names the chain of liable economic operators down to importers, authorised representatives, fulfilment-service providers and certain platforms; Art. 9 compels disclosure of evidence in a defendant's control; and Art. 10 supplies rebuttable presumptions of defect and causation. Liability is channelled to producers, not deployers, making it an ex-post complement to the AI Act's ex-ante duties.

The complexity presumption and the opacity problem

The Directive's signature AI adaptation is evidentiary. Art. 10(4) instructs a national court to presume defectiveness or the causal link where a claimant faces "excessive difficulties, in particular due to technical or scientific complexity," in proving them — a direct response to the black-box character of machine-learning systems whose internal logic claimants cannot reconstruct. This is reinforced by the Art. 9 disclosure duty and the Art. 10(2)(a) adverse inference when a defendant withholds ordered evidence, so opacity itself becomes a litigation cost for producers rather than for victims. The mechanism interlocks with Regulation (EU) 2024/1689: a proven breach of AI Act logging, transparency or risk-management obligations can feed the presumptions. Scholarship on visibility and identification of AI systems ^1,2 underscores why such infrastructural evidence matters — without IDs, logs and monitoring, even a rebuttable presumption struggles for factual purchase.

Agentic systems and post-market mutability

Two provisions confront AI's dynamism. Art. 7(2)(c) makes a product's "ability to continue to learn or acquire new features after it is placed on the market" relevant to assessing defectiveness, abandoning the assumption that a product is frozen at the moment of sale. Art. 11(2) keeps manufacturers liable for defects introduced by software updates or upgrades, and by machine learning, that remain within their control — narrowing the traditional later-defect defence. These rules speak to the governance gap around autonomous and continually-adapting agents, a frontier that legal scholarship treats as distinct from static models: Kolt's agency-law framing ³ and the agent-infrastructure and delegation proposals ^4,5 argue that attribution and remediation require external systems, while multi-agent failure modes ⁶ complicate the "within the manufacturer's control" line that Art. 11(2) draws.

Fault lines, status, and implementation trajectory

The Directive is adopted but not yet in force: it entered into force on 18 November 2024, yet its substantive liability rules apply only to products placed on the market after 9 December 2026 (Art. 2(1)), leaving a transposition and adaptation window for Member States and producers. Its salience grew when the Commission withdrew the separate proposed AI Liability Directive in 2025, so the PLD now carries the principal EU AI-liability load while remaining deliberately silent on ex-ante topics — biometrics, deepfakes and compute fall to Regulation (EU) 2024/1689 and sectoral law. Critics note real fault lines: the Recital 14 carve-out for non-commercial free and open-source software, the producer-not-deployer channelling that may miss harms from how systems are operated, and the limits of compensation as redress. Contestability research ⁷ warns that monetary liability is a thin substitute for meaningful, accessible avenues to challenge automated decisions.

Cross-jurisdiction comparison

How peer instruments treat the topics Revised Product Liability Directive (Directive (EU) 2024/2853) governs.

°= industry self-imposed voluntary framework. Comparing a voluntary code's "governs" tint with a binding regulation's "governs" tint flattens the legal-force distinction; use the instrument-page banner for the operative status of each.

Does this instrument’s approach work? — the social-science evidence

Aggregated over the 3 topics this instrument governs: whether each harm is empirically real, and whether the peer-reviewed evidence shows governance reduces it. The badge is the epistemic status of the evidence— “thin”/“absent” efficacy evidence is itself a finding (the “second silence”). Each epistemic-status label is Policy Window's editorial assessment of the cited evidence base (a structured classification), not a verdict any single source issues.

Of the 3 governed topics with a social-science evidence review, evidence that governance reduces the harm is established for 0, contested for 0, thin for 0, and absent for 3 — for most, no replicated study yet shows this instrument's approach works (the "second silence").

• Agentic AI Governance

  • Is the harm real?evidence: thin

    The capability that agentic governance targets — autonomous multi-step action — is real and rapidly, measurably advancing: METR finds the task length AI agents complete at 50% reliability has doubled roughly every seven months for the past six years (about 50 minutes for frontier 2025 models), and the UK AI Security Institute's first Frontier AI Trends Report (Dec 2025, >30 systems) reports models now finish hour-long software tasks >40% of the time versus <5% in late 2023. The distinct realized HARM from agency (as opposed to the underlying model) is, however, thinly documented: on consequential real-world tasks agents still fail the majority — Gemini 2.5 Pro completed only 30.3% of TheAgentCompany's 175 professional tasks (OpenHands scaffold, project leaderboard) — so the agency-specific harm magnitude is early and context-dependent rather than established at scale.

    Sources: Kwa, West, Becker et al. 2025 (METR; arXiv:2503.14499, 'Measuring AI Ability to Complete Long Tasks'); UK AI Security Institute 2025 (Frontier AI Trends Report, Dec 2025); Xu, Song, Zhou et al. 2024 (TheAgentCompany, arXiv:2412.14161); 30.3% figure per TheAgentCompany leaderboard (OpenHands)

  • Does governance work?evidence: absent

    There is no impact-evaluation evidence that agent-specific governance reduces agentic harm: the operative regimes — the EU GPAI Code of Practice (published July 2025, voluntary/non-binding), the Seoul Frontier AI Safety Commitments (2024, voluntary), and AISI agent evaluations — are 2024-25 vintage and have never been measured against an outcome. The scholarship itself has not settled the contested unit of regulation: Kolt (2025) argues for governing the agentic relationship via principal-agent and agency-law tools, while Chan, Ezell, Kaufmann et al. (2024) propose agent-specific visibility mechanisms (identifiers, real-time monitoring, activity logging) that remain proposal-stage and unevaluated — meaning the field has design proposals but, as with most frontier-AI rules, the evidence that any of them works is absent rather than merely thin.

    Sources: Kolt 2025 ('Governing AI Agents', 101 Notre Dame L. Rev., forthcoming; arXiv:2501.07913); Chan, Ezell, Kaufmann et al. 2024 ('Visibility into AI Agents', ACM FAccT 2024, pp. 958-973; DOI 10.1145/3630106.3658948); EU AI Office 2025 (GPAI Code of Practice, July 2025); Seoul Frontier AI Safety Commitments 2024

• Individual Redress

  • Is the harm real?evidence: established

    The premise behind redress — that affected people lack meaningful recourse against automated decisions — is real, but the flagship instrument is weaker than commonly assumed. Wachter, Mittelstadt & Floridi (2017) show GDPR creates only a limited 'right to be informed,' not a binding 'right to explanation' of specific decisions; and controlled work finds the explanations actually delivered do not measurably improve lay decision accuracy over showing the bare AI prediction (Alufaisan et al. 2021; and a 2022 meta-analysis by Schemmer et al. — screening 393 articles down to 9 in the final analysis — reports 'no effect of explanations on users' performance compared to sole AI predictions,' even though XAI overall had a positive effect). Honest caveat: the legitimacy/dignity value of being heard is empirically well established in the procedural-justice tradition even where outcome accuracy is unchanged, so 'redress fails' depends on which aim is measured.

    Sources: Wachter, Mittelstadt & Floridi 2017 (International Data Privacy Law 7(2):76); Alufaisan, Marusich, Bakdash, Zhou & Kantarcioglu 2021 (Proceedings of the AAAI Conference on AI 35(8):6618); Schemmer, Hemmer, Nitsche, Kühl & Vössing 2022 (AAAI/ACM AIES '22, meta-analysis)

  • Does governance work?evidence: absent

    There is no rigorous impact evaluation showing that mandated redress mechanisms (right-to-explanation, appeal, human-in-the-loop review) actually reduce erroneous or unfair automated decisions — the evidence that the rule works is itself missing. The closest experimental analogues are discouraging: explanations increase humans' acceptance of AI recommendations regardless of correctness (Bansal et al. 2021), and algorithm-in-the-loop oversight can introduce racial disparities and exhibit automation bias rather than reliably catching model errors (Green & Chen 2019). The procedural-justice literature (Tyler 1990; Lind & Tyler 1988) robustly supports a legitimacy and compliance benefit of fair process, but it measures perceived fairness, not reduction of the substantive decision harm redress is meant to cure.

    Sources: Bansal, Wu, Zhou, Fok, Nushi, Kamar, Ribeiro & Weld 2021 (CHI '21); Green & Chen 2019 (Disparate Interactions, ACM FAT* '19); Tyler 1990 (Why People Obey the Law, Yale Univ. Press); Lind & Tyler 1988 (The Social Psychology of Procedural Justice, Plenum Press)

• Transparency Obligations

  • Is the harm real?evidence: contested

    Documentation artifacts (model cards, datasheets) are well-specified as proposals and are genuinely adopted, but the empirical premise that mandated disclosure produces meaningful transparency is contested. Selbst & Barocas (2018) argue inscrutability and non-intuitiveness are distinct problems and that disclosing rules does not resolve the latter, and large-scale audits find documentation is sparsely and unevenly completed: a systematic analysis of 32,111 Hugging Face model cards (Liang et al. 2024) found environmental-impact, limitations and evaluation sections least often filled, and Bhat et al. (2023, 45 practitioners) found a substantial gap between the documentation proposal and actual practice. Honest caveat: the documentation frameworks themselves are real and adopted, so the dispute is about whether disclosure conveys decision-relevant information, not whether the artifacts exist.

    Sources: Selbst & Barocas 2018 (Fordham Law Review 87:1085-1139); Liang et al. 2024 (Nature Machine Intelligence, s42256-024-00857-z, 'Systematic analysis of 32,111 AI model cards'); Bhat et al. 2023 (CHI '23, 'Aspirations and Practice of ML Model Documentation', DOI 10.1145/3544548.3581518); Mitchell et al. 2019 (FAccT, Model Cards for Model Reporting); Gebru et al. 2021 (CACM 64(12):86-92, Datasheets for Datasets)

  • Does governance work?evidence: absent

    There is no rigorous impact evaluation showing that AI transparency mandates (model cards, training-data summaries) measurably reduce bias, misuse or accidents — the central regulatory assumption is empirically untested, partly because flagship mandates like EU AI Act Art. 53(1)(d) GPAI training-data summaries are only subject to AI Office enforcement/verification from 2 August 2026 (the obligation itself began 2 August 2025 for new models). The closest analogue, mandated consumer disclosure, shows small and context-dependent effects: Bollinger, Leslie & Sorensen (2011) found mandatory calorie posting cut average calories per transaction by about 6%, while Loewenstein, Sunstein & Golman (2014) review evidence that disclosure effects are frequently diminished or even reversed by limited attention and often change provider rather than recipient behavior. These are analogues, not AI studies; no study demonstrates that AI transparency disclosure achieves its stated downstream safety aims.

    Sources: Bollinger, Leslie & Sorensen 2011 (AEJ: Economic Policy 3(1):91-128); Loewenstein, Sunstein & Golman 2014 (Annual Review of Economics 6:391-419, 'Disclosure: Psychology Changes Everything'); EU AI Act Art. 53(1)(d) GPAI training-data summary (obligation from 2 Aug 2025; AI Office enforcement from 2 Aug 2026)