A team from Hamburg University of Technology (TUHH) has won the European IP4OS Open Innovation Challenge. The students and their lecturer, Dr. Jonas Bozenhard from the Institute of Ethics in Technology, have now presented their prototype, “ReuseReady”, in Brussels. The tool helps researchers determine whether open datasets can be used to train AI models and, if so, under what conditions.
For Data Science students Niharika Kiran, Praveen Megharaj and Jesvin Varghese, the presentation in Brussels was a particularly significant achievement. The summit was hosted by the Brussels office of the Consejo Superior de Investigaciones Científicas (CSIC), Spain’s national research organisation. During the session “From Principles to Practice”, the TUHH team presented ReuseReady alongside contributions from the University of Luxembourg and IE University Madrid.
Open does not necessarily mean free to use
But what exactly does ReuseReady do? Consider a typical scenario: A doctoral student comes across an open dataset that she would like to use to train her AI model. Determining whether she can do so safely, legally and responsibly is often far from straightforward. Many datasets are subject to licences and terms of use that restrict how they can be used. This is where ReuseReady comes in. Users enter a link or the dataset’s persistent digital identifier, known as a DOI, into the prototype and describe their planned AI project. The system then examines the licensing terms and conditions of use, the reusability of the data and its suitability for AI training, known as AI readiness. It also assesses the dataset against the FAIR principles, evaluating whether the data is findable, accessible, interoperable with other data and systems, and reusable.
Users receive a risk assessment with explanations of the individual risks, along with guidance on what still needs to be checked and recommendations for the next steps. ReuseReady also identifies licensing options for the results users generate when working with the data. The tool follows a clear principle: AI provides information, guidance and possible courses of action, together with their implications. The final decision, however, remains with the human user. “With ReuseReady, we want to demonstrate how this agenda can be translated into practical tools,” explains Megharaj. “Tools that enable researchers to reuse knowledge responsibly and safely.”
On stage in Brussels
In the afternoon, the team continued to the European Parliament. At a workshop organised by the Panel for the Future of Science and Technology (STOA), Members of the European Parliament and international legal experts, among others, discussed the future of intellectual property in the age of Open Science and AI. “The summit showed that Open Science, AI and intellectual property are no longer separate debates,” says Praveen Megharaj. “Together, they form a shared agenda for innovation in Europe.”
“We had the opportunity to learn from experts, exchange ideas and take part in discussions about the future of Open Science, AI and innovation in Europe,” says Niharika Kiran. “Being able to present ReuseReady made the experience even more meaningful. It allowed us to demonstrate how these ideas can be translated into practical solutions.”
From seminar to prototype
The team’s journey to Brussels was made possible by the Open Innovation Challenge organised as part of the EU-funded IP4OS project. The project explores how intellectual property protection and Open Science can reinforce one another. The challenge sought ideas that bring together artificial intelligence, Open Science and the management of intellectual property rights. A compelling concept was sufficient to enter the competition; a finished product was not required. The TUHH team went one step further and submitted a working prototype.
The story of ReuseReady, however, began earlier, in the Master’s seminar “Ethical Innovation and AI”, developed and taught by Bozenhard. “In this practice-oriented course, my students developed innovative solutions to challenges arising from generative AI,” says Bozenhard. “This is where the initial ideas that later became ReuseReady first emerged.” As part of the seminar, the participating students presented their proposed solutions at the International Conference on Ethics Education in Amsterdam. Kiran, Megharaj and Varghese introduced their “Ethical Analyser”, a browser extension designed to assess the credibility of AI-generated responses. Experts at the conference recommended adapting the approach to address intellectual property protection. The team took up the suggestion and, with Bozenhard’s support, fundamentally reworked their original idea.
For Bozenhard, the success in Brussels is above all a testament to his students’ achievements. “It was wonderful to see ideas from my seminar make an impact beyond the lecture hall,” he says. “I am incredibly proud of my students, their outstanding work and their courage in speaking at such a high-profile event.”
