AI Contract Disputes Are on the Rise — Are You Ready?
Artificial intelligence (AI) is rapidly transforming how businesses operate. It is also giving rise to a new category of legal disputes: AI-related contract litigation.
These disputes are common whenever a new technology arrives. They often happen because either the seller or buyer — or both — does not fully understand the technology, or because the technology’s limitations are not properly disclosed.
The numbers tell the story.
A 2024 RAND Corporation study found that more than 80% of AI projects fail — roughly double the failure rate of traditional software projects. A 2025 World Economic Forum report concluded that 88% of companies now use some form of AI in candidate screening alone.
When that much money changes hands on technology that falls short more than 80% of the time, disputes are inevitable.
For Illinois lawyers and businesses, the stakes are especially high. The state’s amended Human Rights Act, effective January 1, 2026, the Biometric Information Privacy Act (BIPA), and the Artificial Intelligence Video Interview Act already give clients more ways to sue — and more ways to be sued.
Why AI Is Creating New Contract Disputes
AI tools are being marketed across nearly every industry — health care, logistics, finance, real estate, manufacturing — as silver bullets to streamline operations and cut costs.
Many do not perform as advertised.
AI is still in its infancy as a technology, and the gap between the sales pitch and operational reality is often wide. The result is predictable: the AI underperforms or causes unforeseen problems.
Common issues include overstated capabilities in contracts or sales materials, misunderstood buyer expectations about what the AI can realistically do, integration failures caused by poor vendor support or miscommunication, undisclosed third-party data usage or security flaws, noncompliance with privacy laws or industry regulations, and delayed or unusable deliverables from AI vendors.
These risks are not entirely new. Any contract can fail when the parties do not clearly define expectations and limitations.
Consider an older parallel: a business hires a phone answering service in the 1920s. If the parties fail to specify operating hours — for example, 9 a.m. to 5 p.m. — the business may assume it is getting round-the-clock coverage. The same dynamic plays out with AI. Unless the technology’s limitations and the parties’ expectations are clearly documented, disputes become far more likely.
What makes AI distinctive is the “black box” problem. Even the vendor sometimes cannot fully explain why a model produced a specific output. That reality makes drafting performance benchmarks and warranties significantly harder than in prior software eras.
It also makes litigation more complex. Lawyers increasingly find themselves arguing about training data, model versions, hallucination rates, and statistical bias — territory most contract attorneys never expected to navigate.
Key Legal Claims in AI Business Disputes
When a client is harmed by a failed AI project, several legal theories may apply.
Breach of Contract
Breach of contract remains the most obvious claim. Common allegations include missed performance benchmarks, missing features, and failure to meet promised accuracy thresholds.
The challenge is that many AI contracts contain vague language such as “industry standard,” “best efforts,” or “commercially reasonable accuracy,” which gives vendors substantial room to argue performance was adequate.
Fraud or Negligent Misrepresentation
Salespeople frequently make claims the product cannot support.
A recent case from British Columbia, Moffatt v. Air Canada, 2024 BCCRT 149, illustrates the issue. After Air Canada’s chatbot incorrectly informed a customer that he could retroactively apply for a bereavement fare discount, the airline later refused to honor the discount and argued the chatbot was effectively a “separate legal entity” responsible for its own statements.
The tribunal rejected that argument, finding Air Canada liable for negligent misrepresentation.
Although the decision is not binding U.S. authority, the reasoning is highly persuasive: businesses are generally responsible for what their AI systems communicate to customers.
Expect U.S. courts to increasingly adopt similar reasoning.
Breach of Warranty
Sophisticated buyers can negotiate AI-specific warranties directly into their contracts. These may include warranties regarding training data quality, accuracy levels, non-infringement, regulatory compliance, and freedom from known bias.
When those warranties are breached, the buyer has a cleaner cause of action.
The problem is that most off-the-shelf AI contracts are drafted heavily in favor of the vendor and contain sweeping disclaimers designed to shift risk to the customer. In many cases, the contract effectively says: “If the AI fails, that is your problem.”
Buyer-side counsel must negotiate aggressively for meaningful warranty protections at the outset.
Data Privacy Violations
AI systems are extraordinarily data-hungry, and many vendors quietly use customer data to train or improve their models.
That practice can implicate HIPAA, the Biometric Information Privacy Act (BIPA), the Illinois Personal Information Protection Act, and an expanding patchwork of state privacy statutes.
BIPA deserves special attention in Illinois practice because its statutory damages and private right of action have already generated some of the largest privacy settlements in the country.
Any AI tool processing facial geometry, fingerprints, voiceprints, or similar biometric identifiers can potentially trigger liability.
Discrimination Claims
Businesses using AI in hiring, lending, insurance, or housing face substantial discrimination risk.
In Mobley v. Workday, Inc., the court allowed claims to proceed against the AI vendor itself — not just the employer — based on allegations that Workday’s AI screening tools discriminated against applicants.
The implications are significant. Vendors cannot simply hide behind disclaimers stating the employer made the “final decision,” and employers cannot blindly rely on vendor assurances.
Copyright and Intellectual Property Claims
Generative AI trained on copyrighted material has already triggered a wave of litigation.
Businesses using generative AI face downstream risk because AI-generated outputs may infringe copyrights, trademarks, or trade dress rights.
Well-drafted AI contracts should allocate these risks through indemnification provisions, and buyer-side counsel should insist on meaningful indemnity language whenever generative AI tools are involved.
Real-World Scenarios
Some recurring fact patterns illustrate how these disputes unfold in practice.
An AI vendor promises 30% cost savings and 90% accuracy. After implementation, savings are marginal and accuracy hovers around 60%. The contract’s vague “best efforts” language makes it difficult to prove a breach.
A marketing agency uses a generative AI platform that produces video content containing unauthorized song lyrics. The record label sues the marketing agency — not the AI vendor — and the agency lacks indemnification protection.
A bank deploys AI underwriting software that disproportionately denies loans in certain ZIP codes, triggering a federal fair-lending investigation.
In Mata v. Avianca, Inc., lawyers submitted a brief generated by OpenAI’s ChatGPT containing six entirely fictional cases. The court imposed sanctions.
AI tools that “hallucinate” can produce very real consequences for the lawyers and businesses relying on them.
The common thread is straightforward: AI is only as good as the data it is trained on, the contract governing its use, and the human oversight supervising its output.
Practical Drafting Tips
Businesses can significantly reduce risk through careful drafting.
Performance obligations should be defined numerically wherever possible. Terms like “commercially reasonable accuracy” are litigation bait. Instead, contracts should define measurable accuracy thresholds, false-positive rates, latency benchmarks, and acceptance-testing criteria.
Intellectual property risk should also be addressed directly. Buyers should demand indemnification for third-party IP claims arising from both training data and AI-generated outputs.
Contracts should clearly state whether customer data may be used to train the vendor’s models, what security standards apply, and whether the data must be deleted when the relationship ends.
Where AI affects consequential decisions — such as employment, lending, insurance, or housing — buyers should retain the ability to audit and explain outputs.
For higher-risk applications, businesses should also maintain a “human in the loop” by requiring human review of important AI-generated decisions and documenting the review process.
Because AI systems evolve over time, contracts should address model drift, retraining obligations, version control, and notice requirements for material model changes.
Termination provisions should also address data return, deletion certifications, and continued access to logs or records needed for future disputes.
Looking Ahead
The legal landscape surrounding AI is moving quickly.
Colorado’s AI Act, New York City Local Law 144, California’s AI bias regulations effective October 1, 2025, and the European Union AI Act all impose expanding compliance obligations on businesses deploying AI systems.
Illinois has also moved aggressively into this area.
Effective January 1, 2026, HB 3773 amends the Illinois Human Rights Act to prohibit discriminatory AI use in employment, while the Illinois Department of Human Rights has issued draft rules concerning notice and recordkeeping obligations.
Lawyers advising AI buyers and sellers must track these developments carefully and draft contracts that anticipate regulatory change rather than merely reacting to it.
AI is not going away, and neither are the disputes it creates. Lawyers and businesses that learn to evaluate AI contracts carefully — with attention to performance standards, privacy obligations, bias risks, and intellectual property exposure — will be significantly better positioned when disputes arise.
If you have questions about a potential or ongoing AI contract dispute, consult experienced counsel who understand both the technology and the evolving legal landscape.
About King & Jones
King & Jones has represented clients in contract interpretation, arbitration, disputes, and litigation for more than 35 years.
The firm has guided businesses through major technological shifts across numerous industries nationwide and continues to advise clients navigating the rapidly evolving risks surrounding artificial intelligence.
This article is for informational purposes only and does not constitute legal advice. Consult a qualified attorney regarding your specific situation.
