Scott Steinberg · Consultant, Analyst and Business Strategist
AI disputes ask what an ordinary person understood they were dealing with: whether a feature did what the LLM maker said, whether someone knew they were talking to a machine, and whether the personalization working on them was ever explained. A strategist and consultant to 3000+ businesses and brands, Scott Steinberg addresses AI as audiences actually meet it.
Expert reports, declarations, deposition and trial testimony on AI feature claims, chatbots and assistants, personalisation and pricing, generated content disclosure, AI in devices and apps, etc.
Sample areas include ML, LLMs, automation and AI products people buy and use. Enterprise procurement and governance questions, and model-level technical questions, also arise from clients.
Marketing for AI has outrun measurement badly. Disputes examine accuracy and capability representations, demonstrations filmed under conditions no user encounters, and terms like smart or intelligent that carry no fixed meaning.
Testimony addresses what an ordinary purchaser would have taken a claim to promise at the time it was made, and what the product actually delivered in normal use.
Conversational products raise questions about whether users knew they were dealing with software, how handoff to a person worked, what the system did with what people told it, and how it behaved outside its competence.
Testimony addresses deployment practice for consumer conversational systems and what the interaction record shows about actual behaviour.
Recommendation and personalization shape what consumers see and buy, and disputes examine what signals drove the system, whether commercial arrangements influenced ranking, and what users were told about any of it.
Testimony addresses how these systems are built commercially and what disclosure practice accompanied them at the time.
Dynamic and individualised pricing raise questions about what inputs set a price, whether comparable customers saw different figures, and whether the practice was disclosed in any meaningful form.
Testimony addresses pricing system practice in the category and what a consumer could observe about how their price was determined.
Where content shown to consumers was machine-generated, disputes examine whether it was labelled, how prominently, and whether the labelling would register with an ordinary viewer in that format.
Expectations here changed quickly under platform and regulatory pressure, so the applicable standard depends heavily on the date.
Devices and apps sold on AI capability raise questions about on-device versus cloud processing, what continues working without connectivity, and what happens when a feature is withdrawn after purchase.
Testimony addresses what a buyer was told about capability and dependency, and how the product's behaviour compared.
AI subscriptions introduce usage caps, model substitution, quality changes between tiers and capability that shifts without notice while the price stays constant.
Testimony addresses what subscribers were told they were buying and how the delivered service compared across the subscription period.
Consumer AI in products used by young people raises questions about audience composition evidence, what defaults applied to younger accounts, and what parents were told about the feature.
Testimony addresses practice for AI features in youth-reaching products and what the operator's own data showed about who was using them.
AI features ship and change continuously, so establishing what the product actually did and disclosed during the relevant window is foundational. Engagements are accepted from either side.
Written opinions on consumer AI claims, disclosure adequacy and what a product's features actually delivered.
Testimony on what an ordinary user would have understood a product to be doing and why.
Responsive analysis of capability assertions, demonstration conditions and consumer understanding claims.
Non-testifying review of product releases, marketing materials and interface history, plus discovery scoping.
Scott Steinberg is a business and technology analyst, strategic consultant and AI trends expert with nearly 30 years of experience providing management and strategic consulting services to more than 3,000 businesses and brands ranging from startups to government agencies and Fortune 500 firms.
He has testified in sample areas including intellectual property — copyrights, trademarks and patents — patent infringement, marketing, branding, video games, mobile applications, consumer product development, and the growth and monetization of online distribution platforms.
He is the author of over 30 books and has published more than five thousand articles addressing areas including but not limited to marketing, technology, leadership, innovation, advertising, digital transformation, data privacy and social networks. He appears regularly on ABC, CBS, CNN and NBC, and has led seminars and training programs for organizations including Wells Fargo, the PGA Tour, Century 21, Ford, Dell and Procter & Gamble.
His consulting work has been broadly recognized. He has served as a thought leader for the American Bar Association and Corporate Counsel magazine, and has received honors from the International Association for Scholastic Excellence, Fortune, and the 21st Century Icon Awards, among others.
AI as consumers actually encounter it: feature and capability claims on products they buy, chatbots and assistants they interact with, personalisation and pricing systems operating on them, and disclosure about generated content. Enterprise procurement and governance, and model-level technical questions, are addressed on separate practices.
By establishing what an ordinary purchaser would have understood the claim to promise at the time it was made, given how the category was being marketed, and comparing that to what the product delivered under normal use rather than under demonstration conditions. Both halves have to be pinned to the same period.
Expectations around labelling generated content and disclosing automated interaction changed rapidly under platform and regulatory pressure. A presentation that was unremarkable in one year sits outside convention two years later, so the applicable standard has to be established for the relevant date rather than assumed from current practice.
Product release and version histories, marketing and packaging materials as published, interface archives, feature flag and rollout records, support tickets reflecting user confusion, and internal capability testing. Consumer AI features change continuously, so establishing what shipped when is foundational.
Yes. Consulting-only engagements cover case assessment, reconstruction of product and marketing history, review of interface and disclosure evidence, deposition question preparation and critique of an opposing expert's capability assertions, without a disclosed report or testimony.
Initial conversations about scope, timing and conflicts are without charge. Helpful detail includes the product and period at issue, the representations involved and any expert disclosure deadline already set.
A conflicts check is run before any substantive discussion. An initial note listing the party names and a one-line description of the dispute is enough to start; please hold privileged or confidential material until the check clears.