Summary of Key Points from the California Attorney General’s Advisory on AI Regulation
At this link you can find the California Attorney General’s Office Advisory on AI. I have summarized the key points below.
(This is yet another example of the content that will soon be available to paid subscribers exclusively).
General Overview
• The advisory outlines how California’s existing laws apply to AI, emphasizing consumer protection, civil rights, competition, and data privacy.
• AI’s potential for innovation comes with risks such as bias, discrimination, fraud, disinformation, and environmental harm.
Legal Obligations for AI Developers and Users
• Transparency: Disclose how AI impacts consumers’ lives and whether consumer data is used for training.
• Accountability: Ensure systems are validated, tested, audited, and comply with laws.
• Risk Mitigation: Prevent harm to individuals, businesses, and the environment.
Key Areas of Law
1. Consumer Protection and False Advertising
• Misleading claims about AI capabilities, accuracy, or performance are prohibited under the Unfair Competition Law (UCL) and False Advertising Law.
• Businesses are liable for deceptive uses of AI, such as deepfakes, impersonations, or misrepresentation of AI-generated content.
• AI tools must not foster fraud or harm competition.
2. Civil Rights
• Discrimination by AI systems violates laws like the Unruh Civil Rights Act and Fair Employment and Housing Act (FEHA).
• Developers and users must prevent bias or discriminatory outcomes, including indirect effects on protected classes.
3. Data Privacy
• California Consumer Privacy Act (CCPA): Regulates the use, sale, and retention of consumer data.
• Special protections apply to sensitive data (e.g., health, education, neural data).
• The California Invasion of Privacy Act (CIPA) prohibits the unauthorized use of communication data and biometric identifiers.
4. Election Integrity
• Prohibits AI-generated deceptive election materials, undeclared chatbots, and impersonation of candidates or initiatives.
• Platforms must label and remove AI-generated election disinformation.
5. Competition Law
• AI systems must not facilitate anticompetitive practices (e.g., price-fixing, market manipulation).
• Large AI companies face scrutiny under antitrust laws for actions harming competition.
New AI-Specific Laws (Effective January 1, 2025)
• Disclosure and Labeling:
• AI training data must be disclosed (AB 2013, SB 942).
• AI-generated content must include visible markers (SB 942).
• Telemarketing calls using AI must disclose AI involvement (AB 2905).
• Unauthorized Use of Likeness:
• Requires consent for digital replicas in entertainment (AB 2602).
• Prohibits nonconsensual use of deceased personalities’ likenesses (AB 1836).
• Healthcare:
• Licensed physicians must supervise AI tools used in medical decision-making (SB 1120).
• AI and Criminal Acts:
• Expands prohibitions on AI-generated child exploitation content (AB 1831, SB 1381).
• Criminalizes nonconsensual deepfake pornography (SB 926).
General Guidance
• Entities must ensure compliance with all relevant laws, including tort, public nuisance, and environmental regulations.
• Use of AI does not exempt businesses from liability under existing laws.