Introduction
Responsible AI is crucial today because AI systems now influence people’s lives, rights, opportunities, safety, and freedoms at a scale and speed never seen before. When AI goes wrong, the impact is no longer theoretical; it is immediate, widespread, and often irreversible.
AI systems today make decisions that directly affect human rights. It is well understood that AI influences decisions around us. From credit, loans, and financial eligibility, healthcare triage and diagnostic support, hiring screening and promotions to policing, sentencing, and surveillance, there have been numerous legal cases of data bias, data discrimination, etc. We also have many legal cases challenging copyright infringement and inappropriate leadership behavior – analogous to “stealing creative works.”
Although we have ethical frameworks with guidelines that are well understood, we still lack the governance with legal teeth to ensure humans act ethically in using AI methods, and if they don’t, there must be consequences. We need to value our privacy, security, and ethical behaviors – versus going rogue and going against “fairness” norms to guide society forward.
We live in a time where we have few internationally aligned laws that govern clarity on responsible AI practices and increasingly, we have fragmentation of diverse laws across the USA, EU, and Asian countries. The EU has a 3-tiered risk structure, and legislation is advancing, albeit implementing the legislation is facing controversy, particularly from the UK.
This article highlights a number of legal cases that provide context of the AI industry’s evolution in ensuring duty of care, privacy and security are being taken seriously by the judicial system, irrespective of not having clearly defined AI legislation in place. Increasing leadership knowledge at board directors and C-Levels is crucial to appreciate that opportunities with AI are great, but also the risks are high when not managing the AI journey lifecycle.
Legal Case Studies Advancing Ethical Rights
Although some of these cases are still working through the legal system, there are increasingly heavy fines for inappropriate AI practices, which gives rise to increased ethical guardrails and legal precedents to give plaintiffs some level of protection, while the legal frameworks continue to dawdle slowly forward, although there are many U.S. states like California, Utah, etc., that have stepped up, recent developments from the Trump administration to curtail the 50 states to have different AI governing laws, a key development to monitor.
Mobley v. Workday (2024–2025). In this case, a plaintiff alleges that Workday’s AI-driven applicant screening software systematically discriminated against older applicants (age > 40) and also unfairly affected people with disabilities or other protected characteristics. In May 2025, a U.S. federal court allowed the class-action lawsuit to proceed. The Big Risk: The case signals major potential liability for AI hiring-screening vendors and employers, under anti-discrimination law (e.g., age, disability, race).
State Farm – Insurance / Claims-Screening Bias: Algorithm case (2023-2024) Plaintiffs allege that State Farm used a machine-learning algorithm to screen for allegedly fraudulent claims, but that algorithm disproportionately flagged claims from Black homeowners, effectively discriminating by proxy (through variables correlated with race). The Big Risk: The case demonstrates that AI bias risk extends beyond “criminal justice or hiring” – to financial services/insurance, where biased algorithmic scoring can harm equal access to services/compensation
Safe Rent Solutions – AI in housing/tenant screening: (Settlement 2024). A class-action suit alleged SafeRent’s algorithmic tenant-screening system discriminated against renters- especially based on race and income (e.g., tenants on housing vouchers), resulting in unfair denials of tenancy. The Big Risk: The case ended in a US$2.2 million+ settlement and a change in the company’s screening algorithm. It’s one of the first major legal precedents where algorithm-based housing decisions triggered systemic-discrimination claims.
Surveillance / Facial Recognition & Biometric AI: Clearview AI and Privacy / Civil-rights lawsuits (2023–2025). Clearview collected billions of photos (scraped from the internet) for its facial-recognition database, sold to law enforcement, raising concerns of wrongful identification, privacy intrusion. Clearview AI has faced multiple fines in different jurisdictions, with some being overturned. Notable fines include a £7.5 million fine from the UK’s Information Commissioner’s Office (ICO), which was initially overturned but is being appealed again by the ICO. The company was also fined €20 million each in France, Greece, and Italy, €30.5 million in the Netherlands, and an earlier UK fine of $9 million was successfully appealed but the ICO has permission to appeal that decision. The Big Risk: This case demonstrated a disproportionate misidentification of people of colour regulators and privacy-advocacy groups in multiple jurisdictions (EU countries, US states) taking action, with growing fines and legal scrutiny illustrating that biased AI + data misuse can trigger both civil rights.
Intuit / HireVue (ACLU Complaint) (March 2025).The ACLU filed a complaint against Intuit and its AI video-interview provider, HireVue, alleging discrimination against a deaf, Indigenous job applicant. The Big Risk: The claim involves violations of the Americans with Disabilities Act (ADA) and other anti-discrimination laws because the AI allegedly penalized non-standard speech patterns used in interviews (especially with facial or voice analysis) can discriminate against people with disabilities or non-normative behavior, leading to legal exposure under civil rights laws.
AI Deepfake Audio (Racist / Antisemitic) (April 2025). A former high school athletics director was sentenced to jail for creating a racist and antisemitic deepfake audio clip using AI, falsely attributing hateful statements to a school principal. The Big Risk: Generative AI misuse for harassment or defamation can lead to criminal or civil liability, especially when used to create false, harmful statements attributed to real people.
Raine v. OpenAI (August 2025). This lawsuit alleges that OpenAI removed safety protocols (that would have stopped self-harm content) from GPT-4’s conversational model, leading to harm. The Big Risk: The claim centers on economic damages, including funeral costs, because the plaintiff says AI failed to properly respond to suicidal ideation. The lack of robust safety guardrails in conversational AI, especially for vulnerable users can lead to negligence claims or wrongful death suits.
Samsung (Biometric / Face Template Collection) (2024). A class-action alleged Samsung’s Gallery app captured and stored users’ biometric face template data without proper consent, violating Illinois’ Biometric Information Privacy Act (BIPA). The Big Risk: The case highlights tension between device-level AI (or on-device recognition) and biometric data regulation. AI leveraging biometric data (face templates, voice) is highly regulated; unauthorized data capture may trigger strict privacy law consequences.
AI Licensing & Content Training Lawsuits Against OpenAI /ChatGPT in EU / Germany (2025). A German court recently ruled that ChatGPT violated copyright law by training on song lyrics without permission. This case is seen as a precedent for restricting how large language models train on copyrighted content. The Big Risk: AI companies may be legally liable if they use copyrighted material without licensing or a proper legal basis for training.
AI Copyright Lawsuit Against Anthropic (2025). Anthropic agreed to pay $1.5 billion US to settle a class-action lawsuit from a group of authors who accused the artificial intelligence company of using pirated copies of their books to train its AI chatbot, Claude, without permission. Anthropic and the plaintiffs in a court filing asked the U.S. District Judge William Alsup to approve the settlement, after announcing the agreement in August without disclosing the terms or amount. “If approved, this landmark settlement will be the largest publicly reported copyright recovery in history, larger than any other copyright class action settlement or any individual copyright case litigated to final judgment,” the plaintiffs said in the filing. The Big Risk: The proposed deal marks the first settlement in a string of lawsuits against tech companies, including OpenAI, Microsoft, and Meta Platforms, over their use of copyrighted material to train generative AI systems. “This historic settlement is a vital step in acknowledging that AI companies cannot simply steal authors’ creative work to build their AI just because they need books to develop quality LLMs,” Authors Guild CEO Mary Rasenberger said in a statement.
AI Copyright Infringement Against Cohere (2025). Even Canadian newest AI Darling, Cohere, a competitor to Anthropic and OpenAI is currently facing a significant copyright and trademark infringement lawsuit filed by a coalition of major news and magazine publishers in the U.S. and Canada. The publishers allege that Cohere used their content without permission to train its AI models and that its AI tools produce verbatim copies or close summaries of their articles, which harms their business. This case has a group of 14 publishers including Condé Nast (owner of The New Yorker, Vogue), The Atlantic, Forbes Media, The Guardian, the Los Angeles Times, Vox Media, and Torstar Corporation (owner of the Toronto Star). The lawsuit was filed in the U.S. District Court for the Southern District of New York in February 2025. Copyright Infringement: Cohere allegedly scraped content from the publishers’ websites, including paywalled material, to train its large language models (LLMs) without a license or compensation. The publishers cite over 4,000 articles they claim were infringed upon.
You may remember at the World Economic Forum in Davos (2024) when CEO Marc Benioff stated that the “training data used by large language models, including those by OpenAI, has been “stolen” or “ripped off” from content creators and media outlets.” Benioff, who also owns Time magazine, voiced his concerns that content from publications like Time and The New York Times was appearing in AI results without permission or fair compensation. If the training data has been stolen, “Benioff asserted, arguing that “the intellectual property and work of content creators are being used without payment”. He highlighted that content from media outlets, including his own, is “surfacing in these results because all the training data has been stolen“. Benioff called for a “standardized system of payments to fairly compensate content creators for their work, a “bridge” he felt the AI companies had not yet crossed.”
Conclusion
These legal cases’ ruling outcomes are positive developments, increasing legal and public awareness of the risks of a faulty or biased AI model that can unjustly deny opportunities, cause direct harm to individuals as well as protect authors’ creative works.
Responsible AI ensures that models are: explainable, auditable, protect individual rights, and reduce bias. It is also important to understand that AI Models that are more complex are harder to interpret. This is particularly true as Large generative models (LLMs) and deep neural networks are not transparent, can easily learn from uncontrolled datasets, can easily produce hallucinations, and embed hidden biases.
Responsible AI frameworks also help organizations understand model behavior, control data lineage, prevent harmful or discriminatory outcomes, and implement “safe-to-fail” guardrails.
Actionable Insight:
Board of Directors, C-Level Executives, CFOs, COOs, and Chief Legal Counsel and Legal/Assurance Advisors have a major role to play to ensure their organizations and customers have a Responsible AI ethical governance and risk operating framework (robust policies, practices and control systems to manage the successful adoption and scaling of AI technologies).
Our company has been actively training and building governance muscle to Board Directors and C-Levels. We have developed an end-to-end AI Advisory Services Offering, and partner with leading organizations, like: Bedford -TRANSEARCH, EY, Kyndyl and SaskTel to support our customers’ digital transformation journeys. We have a great deal to accomplish to advance AI Digital Transformation successfully.
Plus – stay tuned for more information on our New Book, The AI Precipice, An Executive Guide to Balancing AI Innovation and Safety, co-authored with Cathy Cobey, Global AI Assurance Lead, EY (2026 Routledge Book – April 2026) we are providing new frameworks and control guidance to advance our industry and help us future proof AI deployments more successfully.
Notations:
-
See our SalesChoice/EY/BedfordTRANSEARCH white paper series on Future Proofing AI in your organizations
-
See our AI Podcast/Webinar Series, here.
To stay updated about the latest in AI and its impact, subscribe to our CEO’s weekly AI Insights newsletter
