Why do European biotech researchers put AI data mining first for 2030?
An ERC survey of 388 funded researchers ranks AI-enabled life-science data mining above 29 other emerging technologies. It is a useful signal from Europe's research portfolio, not a forecast that AI will deliver clinical or commercial results.
The short version
- What happened
- The ERC analysis covers 2,416 biotechnology projects in 28 countries that received €4.09 billion from 2014 to 2023; machine or deep learning appeared in 237 projects.
- Who it affects
- Patients and consumers should not expect immediate new treatments from this ranking; it describes research priorities rather than validated products.
- What remains unproven
- The survey response rate was 19.6%, so non-response and enthusiasm bias may affect the ranking of emerging technologies.
Keep in mind
The survey response rate was 19.6%, so non-response and enthusiasm bias may affect the ranking of emerging technologies.
Respondents were ERC-funded project leaders, a highly selected European research population rather than a global sample of scientists, clinicians or companies.
The report combines portfolio classification, self-reported expectations and innovation indicators; it does not test clinical benefit, commercial success or causal effects of AI use.
Source analysis · 2 linked sources. Primary analysed source date: 30 September 2026.
Read the full report · 5 min The evidence, methods, implications and limitations
The headline is a research-community expectation, not a technology forecast
The European Research Council's new biotechnology report asks what its funded researchers believe will reshape the field by 2030. AI-enabled life-science data mining ranked first: 85% of survey respondents rated it likely or very likely to have a significant impact, including 59% who selected very likely. That put it ahead of mRNA vaccines at 82%, omics technologies at 78%, tissue engineering at 77% and CRISPR-Cas12 or CRISPR-Cas13 at 76%.
The ranking is informative because it comes from scientists leading frontier-research projects, but it should not be read as a probability that AI will deliver a particular medicine or commercial return. The survey records professional expectations in May 2026. It does not follow products through trials, compare AI-assisted projects with conventional programmes or independently test whether data-mining systems reproduce across laboratories. The correct conclusion is that AI is now embedded in how this research community imagines biotechnology's near future—not that the outcomes are guaranteed.[1][2]
The denominator matters: 388 replies from 1,979 invited researchers
The ERC sent the survey through EUSurvey to all 1,979 grantees leading projects selected for the analysis and received 388 responses, a 19.6% response rate. Respondents rated 30 technologies on a five-point scale, from very likely to highly unlikely, with an additional option for uncertainty. Each technology was rated by 86% to 95% of respondents. Two thirds of respondents came from the ERC life-sciences domain and one third from physical sciences and engineering, broadly matching the wider portfolio; most worked on projects with health applications.
That design gives the findings a defined population and transparent denominator, but non-response remains important. Researchers who see AI as central to biotechnology may have been more motivated to participate, while ERC grantees are not representative of every university, start-up, hospital or country. The 2030 choices were also drawn from OECD and European Commission horizon-scanning work, which makes the list policy-relevant but may frame what respondents considered. Twenty respondents separately raised AI uses in free text, including biological design, foundation models, drug discovery and protein engineering.[1]
AI is already spread through the portfolio rather than confined to an AI category
The wider analysis covers 2,416 ERC biotechnology projects funded between 2014 and 2023 across 28 countries, representing €4.09 billion and 21% of all ERC projects in that period. Machine or deep learning appeared in 237 biotechnology projects, while only 19 were formally classified in the report's AI, Data and ICT sector. The gap suggests AI is often an enabling method inside health, biology or engineering research rather than a standalone research destination. That has practical consequences for funding review, skills and reproducibility: biological teams need data expertise, while model assessment must stay connected to domain evidence.
The report also describes movement towards application, but the pipeline narrows sharply. Biotechnology projects received 531 ERC Proof of Concept grants, and 24 projects went on to receive 25 European Innovation Council Transition grants. Across the portfolio, 211 researchers filed patents and 72 spin-off companies were launched. Those are activity counts, not evidence that the underlying inventions reached patients, scaled manufacturing or produced durable businesses. The ERC itself identifies funding gaps, commercialisation, scale-up, innovation ecosystems and governance as persistent barriers.[1][2]
What institutions can do now—and what would change our assessment
Research institutions can use the finding as a planning signal: improve secure access to life-science data, train mixed biology-and-computing teams, fund independent replication and require clear evaluation against simpler baselines. They should also budget for data stewardship, laboratory validation and regulatory evidence rather than treating model development as the whole innovation pathway. For researchers outside well-funded European networks, shared datasets, open methods and compute access will determine whether AI widens participation or concentrates discovery in a small number of institutions.
Our assessment would strengthen if later studies compare outcomes across AI-intensive and non-AI projects, disclose how many tools progress to laboratory replication, trials or deployment, and reproduce results in less well-resourced regions. It would weaken if the low response rate reflects strong self-selection, if reported AI use remains mostly exploratory or if promising models fail when data sources, laboratories or populations change. The report is a credible map of one major public research portfolio and its researchers' expectations. It is not proof that AI is already the most productive biotechnology investment.[1][2]