Insights
Rights and contracts
229 guides · page 5 of 5
How to find data-use restrictions across hundreds of customer contracts
Contract review for data use restrictions: a search-term list, a triage workflow and a clause inventory to find which customer contracts limit licensing.
How to stop software vendors from training AI on your company data
How to stop vendors training AI on your company data: audit terms, switch off settings, negotiate a no-training clause and keep control of your records.
If AI training is fair use, why do buyers still license data?
Why license data if fair use might apply to AI training? Private records cannot be scraped, and a license gives buyers access, provenance and assurances.
Indemnification in data licenses: who covers which claims
How the indemnification clause in a data license agreement splits IP, privacy, misuse and model-output claims, plus the caps and procedures to negotiate.
Is a company that licenses data a 'data broker' under state laws?
Data broker definitions under state law target sellers of data about people they never dealt with. See when licensing your own records stays outside.
Is an AI data buyer a service provider or a third party under CCPA?
Service provider vs third party under CCPA decides whether licensing records to an AI developer can be a sale. See the role tests, contracts and notices.
Is de-identified data still personal information under state laws?
Is deidentified data personal information? Usually not under US state privacy laws, but only while conditions hold. See what flips it back and what to check.
Labels, annotations and derived data: who owns what the buyer adds?
A derived data clause in a data license decides who owns the labels, embeddings and aggregates a buyer creates. Use this ownership matrix before you sign.
Liability caps in data licensing agreements: typical structures
Limitation of liability in a data license usually has three tiers: a general cap, a super cap and uncapped carve-outs. See how suppliers structure each.
Master data license with dataset schedules: structuring recurring deliveries
How a master data license agreement works with dataset schedules: what goes in the master, what goes per dataset, and how to set up recurring deliveries.
NDA for sharing data samples: clauses to include
An NDA for sharing data samples needs more than standard confidentiality: no training, no retention, a deletion deadline and no re-identification.
No-AI-training clauses in customer contracts: what they cover
A no AI training clause can ban model training, fine-tuning, analytics or aggregated use. See each variant and what it means for licensing records.
Perpetual and irrevocable data licenses: what you give up
A perpetual irrevocable data license can let a buyer use your records with no end date and no exit. See what each word costs and the safer terms to propose.
Purpose limitation: can records collected for one purpose be licensed for AI?
Purpose limitation and AI training: how to test whether records collected for service, support or sales can be reused, and how de-identification changes it.
Red flags in a data license draft: a counsel's checklist
Data license red flags to catch in a buyer's first draft: the phrases that quietly widen use, shift liability or block deletion, and a fix for each one.
Residuals clauses in NDAs: what they mean when AI is involved
A residuals clause in an NDA lets recipients reuse what their people remember. See a sample clause, how AI blurs it, and why to strike it before sharing data.
Right of first refusal and first negotiation in data deals
A right of first refusal in a data license can chill every later bid. Compare ROFR, ROFN and options, and see which terms to define before you grant one.
Right of publicity: employee voices and likeness in licensed recordings
Employee voices and faces in call and meeting recordings raise right of publicity questions in AI licensing. See when transcripts beat audio and what to ban.
Security schedules in data licenses: controls to require from buyers
Security requirements for a data license agreement: twelve controls to require from buyers, from encryption and logging to deletion and incident notice.
Should you delete old data or keep it for AI?
Delete old data or keep it for AI? Apply three tests, legal duty, privacy risk and licensing value, to each record family before an old system is shut off.
Sublicensing and onward transfer in data licenses: affiliates, vendors, clouds
A sublicensing clause in a data license decides who else may touch your records. A checklist for affiliates, annotation vendors and clouds, with flow-downs.
Supplier due diligence questionnaire: what a buyer's counsel will ask
An AI training data supplier due diligence questionnaire covers rights, privacy, provenance and security. Here are the questions and the evidence to prepare.
Supplier quality agreements and sharing NCR and CAPA data
Supplier quality agreement confidentiality decides if NCR and CAPA records can be licensed. Clauses to check and a redaction rule for customer identifiers.
Synthetic data made from your records: who owns it and can it be sold?
Who owns synthetic data generated from licensed data? Why suppliers should treat it as derived data, three contract positions and a clause checklist.
The US Copyright Office report on AI training: what it says about licensing
The Copyright Office AI training report (Part 3, May 2025) treats fair use as case by case and favors voluntary licensing. Five findings for data suppliers.
Third-party content inside licensed records: attachments and embedded documents
Third-party content in a training data license creates warranty risk. A checklist for attachments, vendor manuals, pasted articles and images, and filtering.
Third-party enrichment data in your CRM: can it be licensed?
Third-party data in your CRM usually carries licensing restrictions. How to find enriched fields in Salesforce or HubSpot, strip them and keep your own work.
What happens to a trained model when a data license is terminated?
Data license termination rarely reaches the trained model unless you negotiate it. Compare delete-only, stop-use and retrain terms before signing.
What the 2025 AI fair-use rulings mean for companies licensing data
The 2025 AI fair use rulings in Ross, Bartz and Kadrey were narrow and fact-specific. What they mean for companies licensing private operational data.
Who is liable if a model trained on your data causes harm?
Liability for AI trained on your data usually sits with the developer and deployer. See where a supplier's exposure comes from and the clauses that limit it.
Who owns an AI model trained on your licensed data?
Who owns a model trained on licensed data? Usually the AI developer. See what your company keeps and the license terms that protect it afterward.
Who owns BIM models and shared project data on construction projects?
Who owns the BIM model, RFIs and submittals on a construction project? An ownership matrix by contract form and record type, and what a firm can license.
Who owns customer-written support tickets and chat messages?
Who owns customer support ticket content depends on three separate rights: copyright in the words, your contract license and privacy law. Sort them here.
Who owns end-customer data in distribution and dealer agreements?
Distributor customer data ownership depends on your dealer agreements. See which records you control, which clauses shift rights, and what to check first.
Who owns meeting transcripts and AI notetaker summaries?
Who owns meeting transcripts and AI notetaker summaries: the host company, participants or the vendor? What vendor terms usually say and what to check first.
Who owns telematics and connected equipment data?
Who owns telematics data depends on your provider contract, not on who owns the trucks. Three clauses decide whether you can export, reuse or license it.
Will your company be named in an AI training data disclosure?
An AI training data disclosure may describe or name your company as a supplier. Where names appear, which laws apply and the contract levers you control.