AI Governance and Compliance

AI Governance and Compliance Counsel | New Jersey

AI use policies, vendor and model contracting, training data diligence, disclosure risk and state AI regulation for New Jersey companies.

Most companies do not have an AI problem. They have an AI policy problem

Very few businesses set out to take on AI risk. They accumulate it. A team adopts a tool that was never reviewed. A vendor quietly amends its terms to permit model training on customer inputs. A marketing page describes a rules engine as artificial intelligence. Each decision is small. Together they create exposure that surfaces in a diligence request, a customer audit, a regulator’s inquiry or a lawsuit.

I help companies convert that scattered exposure into something governed, documented and defensible, without building a compliance apparatus that a twelve person company cannot maintain.

A written AI policy is the control most companies are missing

Employees are using AI whether or not the company has approved it. The realistic objective is not prohibition. It is a short written policy that tells people which tools are approved, what information may never be entered into them, who verifies AI-generated work before it leaves the building and who owns the output.

I draft policies that fit the business, along with the pieces that make them operate: employee acknowledgments, onboarding language, handbook provisions and the client-facing disclosures that some industries now expect.

The risk usually lives in the contract, not the technology

AI provisions now appear across software licenses, SaaS subscriptions, vendor agreements, data processing addenda, employment agreements and customer contracts. The terms that matter are rarely labeled clearly.

Contract work in this practice includes:

  • Training data rights and whether a vendor may use your inputs to improve its models
  • Ownership and license terms for AI-generated output
  • Confidentiality, security and subprocessor provisions when data reaches a third-party model
  • Accuracy, human review and acceptable use obligations
  • Indemnification and liability allocation for infringement and erroneous output
  • Audit, notice and change-of-terms rights when the vendor swaps underlying models

What you say about your AI carries enforcement risk

Federal prosecutors and the Securities and Exchange Commission have treated inflated AI claims as securities fraud rather than marketing enthusiasm. The exposure runs through pitch decks, investor updates, product pages, press releases and responses to requests for proposals.

I review AI-related claims for accuracy and substantiation, help companies build the documentation that supports what they say, and advise founders and boards on disclosure before a raise or a sale.

State AI regulation is arriving faster than federal law

New Jersey has multiple AI bills pending, and neighboring states have already enacted requirements affecting hiring tools, consumer-facing systems, disclosure and automated decision-making. Companies operating across state lines increasingly face several overlapping regimes at once.

I track these developments as a member of the New Jersey State Bar Association Artificial Intelligence Committee and translate them into obligations a business can plan around.

Diligence answers the question of what you are really acquiring

Investors, acquirers and enterprise customers evaluating an AI company need to know whether the technology is owned, licensed or borrowed, what data trained it, what rights came with that data and what happens if a key model provider changes its terms.

I conduct and defend AI diligence covering intellectual property chain of title, training data provenance, open source and model license compliance, vendor dependency, customer commitments and regulatory exposure.

Related work

Start with the policy. Then look at the contracts

More on the broader practice at AI Lawyer.

alan@waltercounsel.com • 973-937-8636 • Send a message