Consulting firm notes shift in AI litigation away from issue of how systems are built
According to J.S. Held’s analysis of the emerging artificial intelligence (AI) litigation landscape, technical causation and economic harm frequently rely on a single factual record, which means that coordination among technical, financial, and governance experts is more crucial than ever.
The global consulting firm, headquartered in New York, recently shared the Q3 2026 update to its AI Disputes Monitor, a resource aimed at helping companies understand the legal and business risks arising from AI adoption.
Considering the latest data from its Office of the Chief Intellectual Property Officer, J.S. Held said its tracked dataset now covers 468 disputes, including 115 cases commenced from Jan. 1 to Sept. 30 this year.
J.S. Held confirmed that copyright and training-data disputes continued to be prominent across the dataset. J.S. Held noted that early AI cases tended to centre on which materials trained models and whether developers were authorized to acquire, copy, or utilize protected content.
AI litigation trends
In its news release, J.S. Held acknowledged that legal and intellectual property disputes still comprised the biggest individual market segment in the Q3 2026 update.
Based on the 42 additions as of that quarter, J.S. Held saw a broader litigation landscape with cases increasingly delving into the behaviours of AI systems following deployment, entry into market, and incorporation into consumer, business, and government decision-making, including the systems’
outputs, product safeguards, automated decisions, corporate disclosures, and board oversight.
“The Q3 matters show that deployment risk is becoming a more coherent part of the AI litigation landscape,” said J. Scott Womack, J.S. Held’s senior director at the Office of the Chief Intellectual Property Officer and head of the AI Disputes Monitor data analysis team.
“The litigation is expanding from the inputs used to build AI systems to the consequences of putting those systems into operation,” added James E. Malackowski, chief intellectual property officer at J.S. Held and co-founder of Ocean Tomo, a part of J.S. Held.
Upon assessing the 42 new disputes, J.S. Held pointed to the following AI litigation trends:
- Exposure to AI litigation extended to model developers, technology platforms, companies incorporating AI into their operations, public institutions, corporate officers and directors, and users allegedly breaching safety controls
- The evidence touched on model versions, prompts, outputs, system logs, integration records, testing evidence, data provenance, human-review procedures, governance documentation, market impact, and economic losses
AI disclosure risk
J.S. Held identified two types of AI-related disclosure risk emerging as of the Q3 2026 update:
- statements on whether there has been substantiation of the organization’s advertised AI capabilities, integrations, partnerships, or commercial benefits, commonly associated with “AI washing”
- statements on whether the organization sufficiently revealed the ways AI was disrupting its existing products, pricing models, customer behaviour, or growth prospects
J.S. Held explained that both overstating and understating AI’s capabilities could result in AI disclosure risk.
Integrated expert record
J.S. Held emphasized the importance of an integrated expert record for AI litigation. According to J.S. Held, with the expansion of the AI litigation landscape, the following questions might arise:
- What system version was operating?
- What inputs, prompts, external sources, and integrations impacted the outcome?
- What did the system decide or produce?
- What testing, guardrails, monitoring, and human-review procedures were implemented?
- What did the organization know or represent about the system?
- What alleged harm occurred?
- How should one measure the consequent loss or economic impact?
“The evidence now may need to explain not only where the data came from, but how the system behaved, what controls were in place, what the organization represented about its capabilities, and how the alleged conduct affected economic value,” Malackowski said in J.S. Held’s news release. “These questions cannot be separated neatly into technical, financial, and governance workstreams because they often arise from the same underlying facts.”
“Across product safety, privacy, automated decisions, securities, and governance, a recurring question is whether the organization can reconstruct how the system operated and demonstrate that appropriate controls were functioning. Policies alone will not answer that question,” Womack added. “The relevant evidence may include testing records, system logs, escalation procedures, decision documentation, and proof of meaningful human oversight.”
More international law news on artificial intelligence
Here are some other recent international law news stories revolving around AI.
California Governor Gavin Newsom recently approved Senate Bill No. 574, which prohibits the state’s attorneys from delegating the practice of law to generative AI, but not from using such tools to assist in their legal work.
According to the findings of DISCO’s annual study on legal AI trends among law firms and corporate legal departments, legal teams have been preoccupied with the question of how much legal AI costs, not whether it is accurate.
Enjoy this story? Read the latest international law news on the main page!
