Can a University or Research Lab Sell Its Data to AI Labs?
AITrainingDataBroker.com
Can a University or
Research Lab
Sell Data to AI?
Sell Data to AI?
Who owns it, who can license it, and the checks that come first
Jimmy Lin
Can a University or Research Lab Sell Data to AI?
A university or a research lab sits on years of work that AI labs would love to read, from theses to lab notebooks. Before anyone sells any of it, though, there is a question that stops most deals: who is allowed to license it? There are 3 people with a say and 6 checks to make first, and I will take you through both, starting with a simple scene.
The video
Full script
A university or a research lab sits on years of work that AI labs would love to read, from theses to lab notebooks. Before anyone sells any of it, though, there is a question that stops most deals: who is allowed to license it? There are 3 people with a say and 6 checks to make first, and I will take you through both, starting with a simple scene.
A university research office holds 400 dissertations, and one day an AI lab asks to license them. The author wrote every page, the university gave the degree, and a funder paid for the lab work behind it. So whose is it to license? Most people say the author, because the name is on the cover, and that is only part of the answer. This is general information and not legal advice. First, let me tell you who I am.
I'm Jimmy. I have worked in data at Fortune 500 companies, as a data analyst, a data engineer and a software engineer, and I am trained in data science. I am also a Canadian military veteran, where I worked on data analysis and reporting.
If you want to sell your data, I am your broker and technical advisor. I bring your data to several AI companies, and I guide your team through the technical side, including compliance. You do not pay anything: AI companies pay my fee.
Already shopping your data around? Contact us anyway. We can still work through the terms with you and help you negotiate the best price.
If you are buying data for an AI lab, I find the right data and see it through to delivery, so you can focus on your research.
So why would an AI lab want any of this? Because research has to show its work. A journal article gives you the short version, while a university or lab holds the full thing: theses with the methods and results in full, lab notebooks that record the work day by day, written methods for every step, and instrument logs with the raw readings. That level of detail is hard to find in public writing, and it is what shows a model how research is actually done. But before anyone gets excited, there is a catch: work this rich also raises a question about who it belongs to.
Think of it as 3 people with a say in a single thesis or notebook. The author wrote it, the university may have rules about it, and a funder may have set terms for the work behind it. Which of them holds which right depends on the school's policy, on the funding contract, and sometimes on whether the author wrote it as part of a job. So here is the law that sets the default.
The starting point is US copyright law, section 201. It says the copyright first belongs to the author, so the person who wrote the work owns it. There are exceptions, and the biggest is work written as part of a job, which the US Copyright Office calls work made for hire. Funding can add a layer too, because a law called the Bayh-Dole Act lets a university keep rights to inventions made with federal money, but that covers inventions and not the writing itself. Next, a belief that trips up many owners.
Now, most people think that if a thesis is online, anyone can license it, and that is not true. ProQuest, a company that publishes many dissertations, says in its own author guidance that the author owns and keeps the copyright, and that ProQuest only gets the right to share the work in the ways the author picks, while the author can give the same rights to others. Some theses also carry an embargo, which means a hold on release until a set date. So being able to read a thesis tells you very little about who can license it.
There is one more group to think about, and that is the people inside the research. A thesis built on interviews or patient studies was written under a consent form, and that form says what the participants agreed to. In the US, the federal rule for research on people, called the Common Rule, sets what such a form has to say, and it has a section on broad consent for later research use. But an older form often never mentions AI training, so the question for every thesis is what its participants actually said yes to.
Put that together and you get 6 checks before a thesis goes on any list. First, the author's written yes, then the university's own rules and the funding contract, then any embargo date, then what the participants agreed to, and finally any co-authors, because the copyright law says the authors of a joint work are co-owners. But the list will not hold up unless every one of the 6 is answered in writing, so a thesis that clears all 6 is ready to talk about, and one that does not can simply wait.
Let us go back to the research office with 400 dissertations, and I should say this is an invented example. The office asks every author for a written yes, and it sets aside anything with sponsor limits. After the first pass, 260 are cleared, 90 are waiting for an answer, and 50 are set aside. The AI lab only ever sees the 260, and the other 140 can join later, one answer at a time.
This is not only about universities. If you run a research lab or an R&D department, the same questions come up for your own notebooks, methods, instrument logs and reports. For each one, ask who wrote it and who owns it. Work that employees write as part of the job usually belongs to the company, but contractors, partners and sponsors can change that. Write the answers down now, and you will know your position before any AI lab asks.
Once your checks are done, here is what I do next. I collect offers from several AI companies and AI labs so you can compare them side by side, and you make the final call. On the technical side, I help your team with what to export, the format, and how your data is cleaned. Contracts and insurance are between you and the AI lab and depend on the deal, so we move forward only after your team signs off.
Whether it is a university archive or your own lab or R&D department, the first step is knowing who can say yes.
If you run a company with 30 or more people, the data you already have could become a new revenue stream.
Want to find out?
You can check if yours qualifies, for free, at AITrainingDataBroker.com.
If you have further questions, leave them in the comments, and I will answer them in the next video.
Thank you for watching.
I will see you next time.
More slides
AITrainingDataBroker.com
Martech and Adtech
Company Data
AI Labs Want?
AI Labs Want?
What can be licensed, what is left out, and what to check first
Jimmy Lin
AITrainingDataBroker.com
AI Data Licensing
Liability:
Who Pays?
Who Pays?
Seller, buyer or broker: who answers for what when you license company data to an AI lab
Jimmy Lin
AITrainingDataBroker.com
AI Labs Want
Your CAD Files
Right Now (Oct 27)
Right Now (Oct 27)
An open request for CAD files, what it asks for, and what to do today
Jimmy Lin
