Blog

When AI assists clinical diagnosis, the physician faces a two-sided liability trap — and a black box they may not be able to explain to patients. This post examines the informed consent doctrine, malpractice exposure, and algorithmic bias risks that every health system deploying AI diagnostic tools must address.

AI-powered wearables and remote monitoring platforms face a complex web of legal risks spanning HIPAA, the FTC Health Breach Notification Rule, state consumer health data laws, and unauthorized practice of medicine doctrine. This post maps the compliance architecture every operator in this space needs to build.

The Pennsylvania AG’s lawsuit against Character.AI for unauthorized practice of medicine signals a new enforcement era for health AI platforms. Here’s what the action means for every company building AI systems that touch clinical decisions or coverage determinations.

AI ambient scribes are generating clinical notes at scale — but AI hallucinations, documentation cloning, and systematic upcoding create serious False Claims Act exposure. This post explains the FCA risk landscape and the compliance framework every health system needs before deploying ambient AI.

Pseudonymization is not HIPAA de-identification. AI-generated synthetic data is not automatically outside HIPAA’s scope. This post explains what de-identification actually requires under HIPAA and where common AI healthcare data practices fall short.

HIPAA does not prohibit using patient data for AI-powered population health analytics — but it requires a lawful pathway. This post walks through the four available options: BAA-covered aggregation, shared vendor platforms, de-identification, and the TPO exception, with the real limits of each.

Standard BAAs and SaaS vendor agreements were not built for clinical AI. This post covers the eight contract provisions healthcare organizations must negotiate before deploying AI tools that handle protected health information.

Oregon SB 1546 and Washington HB 2225 take effect January 1, 2027, creating new chatbot disclosure, minor protection, and crisis protocol requirements. Here is what multi-state businesses need to know to comply.

California SB 951 would require 90 days’ advance notice before AI displaces workers, with detailed disclosures and back pay penalties for noncompliance. The causation ambiguity at the bill’s center will generate significant litigation — and the document preservation obligations run backward into the AI procurement phase.

Employees using unauthorized AI notetakers and consumer-tier AI tools create overlapping legal exposures for employers — from recording consent violations and privilege waiver to trade secret compromise and spoliation risk. This post explains each risk and what a compliant AI workplace policy must address.

The DOJ’s 2026 opinion challenging disparate impact theory is the most significant shift in federal civil rights enforcement in a generation — but it does not eliminate employer liability for discriminatory AI hiring tools. The risk has relocated to state agencies and private plaintiffs, not disappeared.

Federal enforcement of AI hiring bias claims has retreated, but legal exposure for employers has not. State laws are expanding, private litigation is accelerating, and the underlying federal statutes remain fully in force.

California’s Civil Rights Council regulations on AI employment discrimination took effect October 1, 2025, creating binding obligations for any employer with five or more California employees using automated decision systems in hiring, promotion, or any employment decision. This post explains who is covered, what bias testing and risk assessments are required, and the seven steps employers must take now to comply.

AI meeting recorders are now standard enterprise tools, but the legal risks — federal and state wiretapping exposure, attorney-client privilege waiver, trade secret leakage, and data retention liability — are far more serious than most organizations recognize. This post covers the full legal landscape and what corporate policies must address.

Allowing a vendor’s AI to train on your customer data is not a routine data processing arrangement. This post covers the intellectual property, privacy law, competitive risk, and contract provisions businesses must address before authorizing any vendor to use customer data for AI model training or fine-tuning.

The EU AI Act, China’s generative AI rules, UK data reforms, and a growing set of US state laws are now creating real enforceable obligations that affect AI technology contracts. This post translates each major regulatory framework into practical contractual guidance for businesses buying, licensing, or deploying AI-enabled services.

When you embed a vendor’s AI technology into your business operations, you build a dependency that can become a serious liability. This post covers vendor lock-in, concentration risk, data portability, transition assistance, and the termination provisions your contracts must include when AI is deeply embedded in your operations.

Most outsourcing contracts contain no mechanism for overseeing vendor AI deployments — no disclosure requirements, no change notifications, no AI-specific audit rights. This post lays out the core contractual provisions needed to maintain visibility, control, and accountability when a supplier’s AI is doing work on your behalf.

AI has broken the FTE-based outsourcing pricing model — vendors are capturing productivity gains while customers pay 2023 prices. This post explains the three-level alternative pricing structure (unit-based, performance-based, and shared risk/reward) and what makes each model legally enforceable.

Vendors are quietly capturing AI-driven cost savings while customers continue paying 2023 prices under legacy outsourcing agreements. This post explains why FTE-based contracts break down when AI automates the work, and what to renegotiate — pricing model, governance, innovation obligations, and termination rights.

Most businesses deploying agentic AI accept vendor SaaS templates that leave critical gaps unaddressed. This post breaks down the five contractual provisions unique to agentic AI — IP ownership, data use rights, human-in-the-loop thresholds, erroneous decision liability, and regulatory change clauses — and what to push for at the negotiating table.

Standard SaaS agreements were designed for software users control — not for autonomous AI agents acting on your behalf. This post explains the structural gaps in traditional SaaS contracts and what a properly constructed agentic AI agreement must address.

A practical legal guide for businesses on protecting brand identity, executive voices, and likenesses from AI-generated replicas — covering state right of publicity laws, the Lanham Act, the TAKE IT DOWN Act, FTC enforcement against deepfake endorsements, and operational steps to defend against voice cloning fraud.

A deep dive into Section 101 patent eligibility for AI and software inventions in 2026 — covering the Alice/Mayo framework, landmark Federal Circuit rulings including Recentive Analytics v. Fox Corp., USPTO examiner guidance, claim drafting strategies, and when to pursue patents versus trade secrets.

A practical guide for businesses on who owns AI-assisted inventions — covering the USPTO’s current inventorship standard, why incorrect inventorship is dangerous, employee and contractor assignment issues, AI vendor terms of service risks, and documentation protocols for protecting R&D output.