What was promised
The Dutch scientific publisher announced the commitment as part of RAISE, the European Commission's pledging initiative for AI in science. The name collides with Elsevier's own Responsible AI for Strengthening Excellence (RAISE) awards for ClinicalKey AI; the pledge described here is the Commission's initiative, not the awards programme.
Elsevier values the package at around €1,000,000 per year, combining direct funding with in-kind technical support. That is Elsevier's own estimate, published in the company's press release on the pledge, and it is not broken down further.
The headline number is the licences. Four thousand of them, split evenly across four products, so 1,000 each. They go to European researchers and clinicians rather than to the public. Elsevier also says it will establish at least three new partnerships with European scientific institutions every year, and support roughly ten PhD candidates, postdocs and master's students annually with training and career help.
If the entire million were spent on licences alone, each one would work out at about €250. That is not what is happening. The pledge also covers technical support, staff time and the partnerships, and Elsevier does not break down how much is cash and how much is services. The €1 million figure is described as an estimate.
Where the licences go
The four tools are LeapSpace, a literature and research assistant; ClinicalKey AI, which supports clinical decision-making; ClinicalKey Nursing AI; and Complete Anatomy, a 3D anatomy platform used heavily in medical teaching.
That mix matters for who benefits. Two of the four are clinical products, so part of the commitment lands in hospitals and nursing education rather than in university labs. The anatomy tool is the kind of software a first-year medical student in Brno or Porto might open on a tablet during a dissection class. Until now, access to products like these has usually been paid for by the institution, department by department.
The Commission is behind the mechanism
RAISE was launched at the European AI in Science Summit in Dublin in October 2026, with more than 1,000 representatives from business, technology, research and policy in the room. The idea is to pull private and philanthropic money into European research infrastructure instead of leaving it to national budgets alone. Elsevier is one of the launch partners listed in the initiative's first wave of commitments, as set out on the Commission's RAISE page.
Brussels has been pushing the same direction for a while: the EU wants European science to use AI without renting all of its compute and software from American cloud providers. The Commission's digital strategy unit has funded AI Factories across member states for exactly this reason.
Amsterdam centre and the AIRIS project
Two specific commitments sit underneath the numbers. The first is a new Centre on Hybrid Intelligence for Trustworthy Research, set up with the University of Amsterdam and Vrije Universiteit Amsterdam. The second is Elsevier's role in AIRIS, a Horizon-funded project working on AI-assisted hypothesis generation in biomedicine.
Hypothesis generation is a narrow, technical field. In practice it means software that reads a large body of published papers and proposes connections a human researcher might have missed. Whether that produces useful biology or a pile of plausible-looking noise is exactly the question the project is meant to test.
What European hospitals still have to check
Software does not arrive in a European clinic free of obligations. Under Regulation (EU) 2026/1744 — the Digital Omnibus on AI, adopted on 8 July 2026, published in the Official Journal on 24 July 2026 and in force since 27 July 2026 — high-risk obligations now apply from 2 December 2027 for stand-alone AI systems and 2 August 2028 for AI embedded in regulated products. The earlier August 2026 date no longer applies.
That is a delay, not a holiday. Transparency duties under Article 50 have been binding since 2 August 2026, with direct enforcement by the EU AI Office and national market surveillance authorities. General-purpose AI models also face mandatory disclosure rules now. And GDPR applies to anything touching patient data, whatever the licence says.
So a hospital taking up a free ClinicalKey AI licence still has to know what the system does, what it does not do, and who carries responsibility when a suggestion is wrong. Free does not mean unregulated.
What the pledge does not settle
Four thousand licences is a pilot, not a redistribution. Europe has hundreds of thousands of active researchers and clinicians. Ten supported students a year is smaller still. The announcement does not describe an open application process, and the allocation logic appears to run through RAISE and the partner institutions rather than through a public sign-up.
Elsevier has been building AI features into its research and clinical products for several years, so part of this pledge is deployment of tools that already existed. That does not make the funding meaningless. It does mean the €1 million covers a mixture of new money, donated access and services the company was already developing.
Can an individual researcher apply for one of the 4,000 licences?
The announcement does not describe an open application route. The licences are tied to RAISE and to institutional partners, so the practical path runs through your university, hospital or research group rather than through a personal request.
Does this change what my institution pays for ClinicalKey AI?
No. The free allocation covers 1,000 ClinicalKey AI licences within this programme. Standard commercial contracts for other institutions are unaffected, and the announcement does not mention any price change for existing customers.
Is the €1 million new spending?
Elsevier describes it as the estimated annual value of direct funding plus in-kind support, and does not separate the two. Some of it is cash, some of it is licence value, technical help and staff time.
