AI at the
Springer Nature
Group
2024
2
Agenda
1. Quick introduction to AI
2. Our approach towards AI
3. AI Innovation @ SN
3
What is AI?
OUR APPROACH TOWARDS AI
Adopting an ethically-focused approach for designing, developing, deploying and using AI solutions
“... this technology will
enhance us.
So instead of artificial
intelligence, I think we'll
augment our intelligence.”
Ginni Rometty
Former executive chair of IBM
4
5
Generative AI
6
AI Does Not
Revolve
Around
Generative AI
Research communities are early adopters of technology!
Internet itself created by scientists for scientists
“The web was originally conceived and developed [at CERN] to meet the demand for
automated information-sharing between scientists in universities and institutes
around the world”
Recent survey by Nature of >1600 researchers shows a community already
embracing AI “AI has enabled me to
make progress in
answering biological
● >50% believe AI tools to become very important or essential questions where progress
was previously infeasible”
● 30% already use AI in their research
● 48% study AI Computational biologist,
Duke University
● only 22% did not use AI yet
And being used across disciplines - from producing new protein structures,
7 improving weather forecasts, to suggesting medical diagnoses
8
AI opportunities outweigh threats
Opportunities to make research faster; however, need to tackle ethical and legal concerns
OPPORTUNITIES THREATS
1. More than 50% believe AI tools to become very 1. 69% believe AI tools can lead to more reliance on
important or essential pattern recognition without understanding
2. 66% believe AI provides faster ways to process data 2. 58% believe results can entrench bias or discrimination
3. 58% believe it speeds up computions in data
4. 55% believe it saves scientists time and money 3. 55% believe tools could make fraud easier
4. 53% believe that ill-considered use can lead to
“AI enabled me to make irreproducible research
progress in answering “AI is challenging
biological questions our existing
where progress was standards for
previously infeasible” proof and truth”
Computational biologist, Image Analyst,
Duke University Jackson Laboratory
Usage?
Recent survey by Nature of >1600 researchers
9
Our approach towards AI
10
We are approaching AI responsibly
Starting cautiously, building capacity while taking into account legal, integrity, reputational risk
Dedicated AI team + training teams across the business to execute the
Building teams
opportunities and make everybody AI savvy
Building
Working with expert suppliers
partnerships
Creation of 5 principles covering dignity, respect, minimising harm,
Awareness of
fairness, transparency, accountability, privacy and data governance to
ethics
govern AI initiatives
Innovation AI governance process to operationalise our principles. Puts principles
process into practice via several decision forums
10
11
How we’re managing risk
Ethics Governance
12
5 principles for AI at Springer Nature
Dignity, Respect and We prioritize human well-being and dignity, and take steps
Minimising Harm to prevent harm to society and the environment.
Fairness and Equity We mitigate the potential for structural bias and inequities.
We disclose when an AI system is being used and explain
Transparency
our processes in accessible language.
We maintain human oversight of the development and
Accountability
outcomes generated by our AI tools and solutions.
Privacy and Data We safeguard personal privacy and follow all relevant data
Governance protection laws.
13
5 principles for AI at Springer Nature
https://group.springernature.com/gp/group/taking-responsibility/ai-principles
14
AI Innovation @ SN
15
3 Innovation Gates
Stop or
1 Promising? 2 Compliant? 3 grow?
Assess innovation Architecture, Assess commercial
and potential technology, legal and and strategic fit
impact ethics checks
Sandbox - Specify idea and impact In production
CUSTOMER VALUE - AI IS TRANSFORMING THE PUBLICATION PROCESS
We focus and structure our ambition across six areas plus the necessary infrastructure
Research Validation & Distribution Discovery
CONDUCT WRITE SUBMIT PEER ACCEPT PUBLISH DISCOVER &
RESEARCH MANUSCRIPT PAPER REVIEW ARTICLE ARTICLE USE CONTENT
Personalize Ensure
Ease Discovery
Recommendation Research Integrity
Enhance Peer Convert content
Review “by a click”
16
17
CONTENT GENERATED WITH AUGMENTED INSIGHTS - PAPER2BUZZ
New content items on Nature Regional Sites, X and SciAm
18
Automatic Journal Recommendations and
Automatic Detection of Manipulated Images
Automatic Transfer Automatic Detection of Duplicated Microscopic Images
Recommendations to Authors in Figure Panels
Thanks a lot
