Our aihorizon R&D team spent two intensive and inspiring days in Munich, advancing the frontiers of Responsible AI through hands-on collaboration, interdisciplinary exchange, and joint planning for future impact.

 

Day One: At Microsoft Munich with Tech for Social Impact

We kicked off the visit with a full-day workshop at Microsoft Munich, working closely with the Tech for Social Impact team. Our shared mission: advancing responsible, ethical, and applied AI solutions that are not only technically sound but socially relevant and inclusive.

Together, we explored working prototypes, research-driven frameworks, and real-world applications, all centered around fostering transparency, safety, and accessibility in AI.

 

Key Projects on the Agenda

 

Horizon App v3 – Previewed new features of our Responsible AI Learning Space, developed to empower educators, NGOs, and policymakers with engaging, interactive tools for critical AI literacy. (Partner: Microsoft Tech for Social Impact)

 

Censorship Project – Co-developing methods to detect and quantify censorship in LLMs, especially in restricted information environments. (Partner: QuantPi & Microsoft Tech for Social Impact – Epistemic Agency Initiative)

 

TerraCore – Advancing our Sustainability Module for real-time CO₂ tracking on Snapdragon devices, with local AI deployment to ensure privacy and performance. (Partner: Qualcomm)

 

Multidimensional Content Safety Framework – Designing a scalable system that balances ethical, legal, and cultural dimensions of content moderation through explainable pipelines. (Partner: Microsoft)

 

AI HELPS.family – Supporting immigrant families with AI-powered translation and document assistance, reducing the emotional and cognitive load of child language brokering. (Partner: Microsoft Tech for Social Impact & Hochschule Darmstadt)

 

In the afternoon, we dove into Design Thinking to sharpen our strategies around:

• Scaling and outreach

• Communication and PR

• Emerging features and next-stage use cases

 

Day Two: Research Exchange at LMU Munich

The second day took us to Ludwig-Maximilians-Universität München (LMU) for a focused and productive exchange with researchers from the Data Science and AI Lab.

The session centered around three key areas of shared interest:

• Our CO₂ tracking initiative (TerraCore) and the potential for climate-tech collaboration

• The Censorship Project, particularly methodological challenges in detecting content suppression in LLMs

• Our Multidimensional Content Safety Framework, where we discussed tunable moderation systems and explainability in content scoring

 

In turn, members of the LMU lab shared insights into their own research on bias mitigation, and applied data ethics — sparking a rich, interdisciplinary dialogue.

The meeting laid the groundwork for concrete follow-up: from joint research opportunities to shared development goals across responsible AI and sustainability domains.