Societal Issues
Whitepaper - Law: AI Societal Issues
Introduction
Artificial intelligence has already transformed society, offering immense benefits in efficiency, innovation, and problem-solving across sectors. But, it also raises profound societal issues that intersect with law, technology, and business. These challenges include privacy erosion, algorithmic bias, job displacement, liability for AI-driven harms, and the need for robust global governance. Drawing from the principles of The Institute for Ethical AI, this whitepaper explores current topics and areas where law needs to evolve to address AI's societal impacts. Looking forward, in 2026, with rapid advancements in AI technologies and increasing regulatory activity, stakeholders involved with legal aspects of AI need to be aware of the fast growing body of case law that balances innovation with ethical safeguards.
Key Societal Issues
Privacy and Data Protection
AI systems rely on vast datasets, often including personal information, which heightens risks of privacy breaches and surveillance. Current laws, such as the EU's General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA), set standards for data handling, but AI's ability to infer sensitive details from seemingly innocuous data demands stronger protections. In the U.S., state-level AI laws effective in 2026 emphasize transparency in data use for AI training, while federal efforts like executive orders signal a push for national standards. Businesses face compliance challenges, as non-adherence can lead to hefty fines and reputational damage. Technologically, privacy-enhancing tools like federated learning help mitigate risks, but legal frameworks must adapt to AI's global data flows to prevent exploitation.
Bias and Fairness
Algorithmic bias in AI can perpetuate discrimination in areas like hiring, lending, and criminal justice, amplifying societal inequalities. For instance, biased training data may disadvantage marginalized groups, leading to unfair outcomes. Legal developments in 2025-2026 include over 1,100 state bills in the U.S. addressing AI fairness, with many requiring bias audits for high-risk systems. The EU AI Act, fully operational by 2026, categorizes AI applications by risk level, mandating rigorous assessments for biased systems. From a business perspective, companies must integrate fairness metrics into AI development to avoid litigation, while technology advances like debiasing algorithms offer solutions. Societally, addressing bias is essential for maintaining public trust and ensuring AI benefits all.
Employment and Economic Impacts
AI automation threatens job displacement, particularly in routine tasks, potentially widening economic gaps. Projections indicate millions of jobs at risk, prompting discussions on universal basic income and reskilling programs. Legally, this intersects with labor laws, where 2026 regulations may require impact assessments for AI deployments affecting workers. Businesses view AI as a productivity booster, with enterprise adoption rising in 2025, but they must navigate ethical obligations to support workforce transitions. Technologically, AI can create new roles in data science and ethics oversight, yet societal issues like mental health from human-AI interactions demand legal protections for equitable economic outcomes.
Liability and Accountability
Who bears responsibility when AI causes harm, such as in autonomous vehicles or medical diagnostics? Traditional liability laws struggle with AI's opacity, leading to a surge in tort and product liability claims. In 2026, U.S. states and the proposed RAISE Act¹ introduce accountability measures, requiring developers to disclose AI decision-making processes. Businesses must mitigate risks through governance frameworks, while technology emphasizes explainable AI to clarify accountability. Societally, this ensures victims have recourse, preventing unchecked AI proliferation.
Intellectual Property
Disputes over copyrights for works created by tools like generative AI spiked in 2025. Legal forecasts for 2026 predict domain-specific AI in law firms to handle these cases, alongside reforms to patent systems for AI inventions. Businesses leveraging AI for innovation must secure IP rights, often through trade secrets, while technology blurs lines between human and machine creativity. This area underscores that AI-generated content challenges IP laws, with need for updated laws to protect creators.
Global Governance
AI's borderless nature requires international coordination. The U.S. Department of State highlights AI's societal advances, but warns of risks like inequality. In 2026, global frameworks like UNESCO's AI ethics recommendations² influence national policies, with the CREATE AI Act³ promoting U.S. experimentation while addressing harms. Businesses operating cross-border face varying compliance, necessitating adaptable strategies. Technologically, standards for AI safety foster collaboration, but societal issues like AI in warfare demand binding treaties.
Intersections with Technology and Business
AI technology mimics intelligence without sentience, integrating with legacy systems to drive business value, yet it poses ethical hurdles like transparency in hybrid environments. Businesses prioritize compliance amid 2026's regulatory surge, investing in responsible AI to mitigate risks and capture opportunities. Law serves as the bridge, enforcing accountability while enabling innovation. For example, in healthcare and justice, AI's societal impacts require legal oversight to balance efficiency with human rights.
Recommendations
To address these issues, stakeholders should adopt responsible AI practices: conduct regular audits for bias and privacy, integrate ethical training in business operations, and advocate for harmonized global laws. As emphasized by The Institute for Responsible AI, aligning AI with human values minimizes harm and maximizes societal benefits, ensuring a future where technology serves all.
DISCUSSION AND CASE LAW
The societal issues posed by AI necessitate ongoing legal scrutiny, as courts worldwide grapple with balancing technological innovation against protections for privacy, fairness, and accountability. This section examines key real-world case law across the identified areas, drawing on established precedents and recent developments up to 2026. These cases illustrate how existing laws are being applied to AI challenges, highlighting gaps that may require legislative evolution. Discussions focus on their implications for technology deployment, business practices, and societal equity.
Privacy and Data Protection
AI's data-intensive nature has led to litigation emphasizing the need for consent and transparency in data usage. A notable case is Carpenter v. United States, 585 U.S. 296 (2018), where the U.S. Supreme Court ruled that accessing cellphone location data without a warrant violates the Fourth Amendment. While not AI-specific, this precedent has influenced AI privacy disputes, such as those involving facial recognition, by underscoring expectations of privacy in digital tracking.
In the AI context, emerging suits like those against Clearview AI for scraping biometric data without consent⁴ have resulted in settlements and injunctions, reinforcing the application of laws like Illinois' Biometric Information Privacy Act (BIPA) (740 ILCS 14/1 et seq.).
These cases highlight business risks, as companies must now implement robust data governance to avoid class-action liabilities, while technologically, they push for privacy-by-design in AI systems. Societally, they address surveillance concerns, ensuring AI does not erode individual autonomy.
Bias and Fairness
Court cases on AI bias underscore discrimination risks in algorithmic decision-making. In Mobley v. Workday, Inc., No. 23-cv-00583 (N.D. Cal. 2023, ongoing into 2026), a federal court granted conditional certification for claims under the Age Discrimination in Employment Act (ADEA), alleging Workday's AI hiring tools disproportionately rejected older applicants.
This landmark ruling signals that AI vendors can be held liable as "employment agencies" under Title VII, prompting businesses to conduct bias audits. Another key case involves Sirius XM, where a 2025 lawsuit⁵ accused the company of using AI in hiring that violated federal anti-discrimination laws.
These precedents emphasize the need for explainable AI to mitigate disparate impacts, with societal implications for reducing inequality in access to opportunities.
Employment and Economic Impacts
AI's role in workforce decisions has sparked cases on job displacement and fair hiring. Building on bias litigation, Kistler v. Eightfold AI (Cal. Super. Ct. 2026) involves job seekers alleging the AI platform collected sensitive data and scored applicants unfairly under the Fair Credit Reporting Act (FCRA). The case highlights transparency issues in AI scoring, urging businesses to disclose methodologies. In Mobley v. Workday, Inc. (as above), the court's allowance of collective action claims addresses broader employment impacts, including automation's potential to displace workers. These rulings encourage reskilling initiatives and impact assessments, mitigating economic divides while fostering AI's productive use in business.
Liability and Accountability
Determining fault in AI harms remains contentious. In Sewell v. Google LLC and Character.AI (settled 2026)⁶, the companies resolved a lawsuit tied to a teen's suicide allegedly influenced by an AI chatbot, agreeing to safety enhancements without admitting liability. This case illustrates product liability principles applied to AI, emphasizing safeguards against harmful outputs.
Another development is in autonomous systems, where cases like the 2018 Uber self-driving vehicle fatality⁷ led to settlements and influenced standards for AI accountability.
Courts are increasingly requiring explainability, impacting businesses by necessitating risk management frameworks and technologically advancing interpretable models to ensure societal trust.
Intellectual Property
AI-generated content has fueled IP disputes. In Andersen v. Stability AI Ltd., No. 23-cv-00201 (N.D. Cal. 2023, trial set for 2026), artists alleged copyright infringement from AI training on their works, with the court allowing claims to proceed into discovery. This tests fair use doctrines, potentially requiring businesses to license data for AI models.
Similarly, The New York Times Co. v. Microsoft Corp. and OpenAI (S.D.N.Y. 2023, ongoing) claims unauthorized use of articles for AI training, seeking damages and injunctions.
The 2025 settlement in Bartz v. Anthropic⁸ for $1.5 billion set a precedent for compensating creators.
These cases clarify IP boundaries, encouraging ethical data practices in business and innovation.
Global Governance
International cases on AI are emerging, often in human rights contexts. While specific court rulings are limited, frameworks like the EU AI Act have influenced disputes, such as potential challenges to high-risk AI under GDPR.
In the U.S., governance intersects with cases like those above, but globally, UNESCO ethics guidelines inform proceedings. For instance, Canadian courts are applying the Artificial Intelligence and Data Act⁹ (AIDA) in oversight cases, emphasizing accountability.
These developments promote harmonized standards, aiding cross-border businesses while addressing societal risks like AI in warfare or inequality.
FOOTNOTES
Footnote 1: New York’s Responsible AI Safety and Education (RAISE) Act, effective 2027, mandates that developers of high-impact "frontier" models implement rigorous safety protocols and report critical incidents within 72 hours. This landmark legislation emphasizes transparency and accountability, aligning state-level oversight with global governance standards to mitigate catastrophic systemic risks.
Footnote 2: Adopted on November 23, 2021, the United Nations Educational, Scientific and Cultural Organization (UNESCO) Recommendation on the Ethics of Artificial Intelligence (AI) serves as the first global standard-setting framework for ethical digital governance.
Though effective upon adoption as a non-binding normative instrument, it mandates that 193 member states implement legislative actions protecting human rights and dignity. This comprehensive directive promotes transparency and accountability while expressly prohibiting AI applications for mass surveillance or social scoring.
Footnote 3: The Creating Resources for Every American To Experiment with Artificial Intelligence (CREATE AI) Act was introduced by a bipartisan group of legislators, including Senators Martin Heinrich, Todd Young, Cory Booker, and Mike Rounds. This legislation establishes the National Artificial Intelligence Research Resource (NAIRR) to provide researchers and students with the computational power and data necessary to develop safe, trustworthy artificial intelligence (AI).
Footnote 4:
Case Citation: ACLU v. Clearview AI, Inc., No. 2020-CH-04353 (Ill. Cir. Ct. Cook Cty. May 11, 2022) (Consent Order)
Federal MDL Citation: In re Clearview AI, Inc., Consumer Privacy Litigation, No. 1:21-cv-00135 (N.D. Ill. 2025) (granting final approval of class action settlement regarding BIPA violations)
Footnote 5:
Harper v. Sirius XM Radio, LLC, No. 2:25-cv-12403 (E.D. Mich. filed Aug. 4, 2025).
Footnote 6:
Garcia v. Character Technologies, Inc., et al., No. 6:24-cv-01903-ACC-DCI (M.D. Fla. Jan. 7, 2026)
Footnote 7: Nat’l Transp. Safety Bd., Highway Accident Report: Collision Between Vehicle Controlled by Developmental Automated Driving System and Pedestrian, Tempe, Arizona, March 18, 2018, NTSB/HAR-19/03 (2019).
Footnote 8: Bartz v. Anthropic PBC, No. 3:24-cv-05417 (N.D. Cal. filed Aug. 19, 2024; settlement preliminarily approved Sept. 25, 2025).
Footnote 9: The Artificial Intelligence and Data Act (AIDA) is currently proposed federal legislation in Canada, introduced as Part 3 of Bill C-27, the Digital Charter Implementation Act, 2022.
As of early 2026, the AIDA has not yet received Royal Assent and is not yet in force. If passed, it will establish Canada's first regulatory framework for the responsible development and deployment of "high-impact" AI systems.
