Open, Ethical, European

Artificial intelligence is reshaping our daily lives – from healthcare and education to climate research and social media. But technological progress must be built on trust. That means dependable data, transparent methodologies, and outcomes that meet ethical standards. Trustworthy AI isn’t just about performance – it’s about fairness, accountability, and clarity.

Twitter/X – A Lesson in Dependency

A cautionary tale comes from the platform formerly known as Twitter, now X. For years, researchers relied on its open APIs to study political discourse, misinformation, and crisis communication. Thousands of studies were based on this data. But when access was abruptly cut off in 2023, many projects were left incomplete, and findings could no longer be verified. This highlights a critical issue: research that depends on private platforms rests on shaky ground. Trustworthy AI requires permanent, independent, and open data infrastructures.

Europe’s Responsibility: Sovereignty Through Its Own Infrastructure

“The reliance on commercial gatekeepers threatens not only scientific progress but also Europe’s digital sovereignty,” says Dr Marco Fisichella, a researcher in trustworthy AI at the L3S Research Center. Europe needs its own infrastructures that uphold ethics, transparency, and data protection – in line with regulations such as the GDPR. Initiatives like SoBigData are leading the way, offering open and equitable access to diverse datasets while embedding fairness, accountability, confidentiality, and transparency into research practices.

From Ethical Principles to Reproducibility

Reproducibility is a cornerstone of scientific integrity. When data disappears or access conditions change unexpectedly, AI results can’t be verified or built upon. “Public infrastructures like SoBigData provide enduring, well-documented datasets and tools that foster reliability and trust,” Fisichella explains. “At the same time, they integrate ethical guidelines to ensure innovation respects privacy and data protection.”

Tackling Societal Challenges Together

Trustworthy AI is more than just technology – it’s a tool for addressing some of the most pressing issues of our time: misinformation, climate change, disaster response, and public health. This demands large, diverse datasets – and a commitment to using them responsibly. “Open access, ethical principles, and advanced AI methods make infrastructures like SoBigData key technologies for AI that serves society,” says Fisichella. Germany plays a pivotal role in this effort.

The L3S Contribution

The L3S Research Center brings together expertise in machine learning, natural language processing, data management, and privacy technologies to develop trustworthy AI systems. One example is the long-term archiving of web and social media data, ensuring scientific results remain reproducible even when commercial platforms change their policies. Another area is federated learning, where AI models are trained on distributed data without compromising privacy. Fisichella is convinced: “By embedding ethical principles, transparency, and reproducibility into its infrastructures, L3S contributes to a European model of AI that is innovative, responsible, and trustworthy.”