Uploaded December 2025 | Updated September 2026, 2 weeks ago
Trustworthy & Sustainable AI: Powering the Future with Next-Generation Computing | Gitta Kutyniok | Ludwig-Maximilians-Universität München
Artificial Intelligence is driving a new industrial revolution—transforming science, industry, communication, medicine, and every aspect of modern life. But as AI systems grow in scale and influence, two fundamental challenges stand in the way of a reliable, long-term future: trustworthiness and sustainability.
Follow SAI Conferences on Linkedin: linkedin.com/company/saiconference
Conference Website: saiconference.com/FTC
In this visionary keynote, Prof. Gitta Kutyniok explores how addressing these challenges requires going beyond algorithms and software—and taking a foundational look at the limits of current computing architectures. Through a deep, yet accessible, journey across mathematics, machine learning, hardware design, and neuromorphic computing, she explains why the next breakthroughs in AI may come from rethinking the very nature of computation.
🔍 What You’ll Learn in This Talk
1. Why Trustworthy AI Is Still an Unsolved Challenge
Safety, reliability, and security risks in modern AI systems
Black-box decisions, bias, privacy violations, and the global trust deficit
The mathematical foundations behind reliability—and what theoretical guarantees can (and cannot) offer
How generalization, expressivity, and explainability drive trustworthy AI
2. The Sustainability Crisis in AI
Why AI’s energy consumption is nearing global production limits
The hidden cost of training, communication, and large-scale inference
Insights from the U.S. decadal roadmap on semiconductor and AI energy consumption
Why simply scaling hardware is no longer viable
3. The Limits of Digital Computing
How traditional digital architectures contribute to reliability and energy challenges
What computability theory tells us about what neural networks can and cannot compute
Why certain inverse problems remain fundamentally non-computable on digital hardware
4. The Case for Next-Generation Computing
How analog, neuromorphic, and brain-inspired chips can overcome foundational barriers
Real examples of neuromorphic hardware (e.g., SpiNNaker) and its energy advantages
Spiking Neural Networks (SNNs): mathematical perspective, biological realism, and efficiency
Why time as a computational dimension opens entirely new AI capabilities
The promise (and limits) of emerging paradigms like optical computing, quantum computing, and bio-computing
5. Building a Future of Reliable & Green AI
Inside Bavaria’s state-level initiative Next Generation AI Computing
The importance of co-evolving hardware and software, rather than designing them separately
Applications in robotics, communication, medicine, and mission-critical systems
A vision for AI systems that are simultaneously powerful, trustworthy, and energy-efficient
🎤 About the Speaker: Prof. Gitta Kutyniok
Prof. Gitta Kutyniok holds the Bavarian AI Chair for Mathematical Foundations of Artificial Intelligence at the Ludwig-Maximilians-Universität München (LMU). She is affiliated with the DLR – German Aerospace Center and the University of Tromsø, and is a globally respected leader at the intersection of mathematics, AI, deep learning theory, and next-generation computing.
Her distinguished career includes:
Founder of EcoLogic Computing GmbH
Member of the Berlin-Brandenburg Academy of Sciences and Humanities and the European Academy of Sciences
Fellow of SIAM (2019) and IEEE (2024)
Plenary and invited lectures at ICM 2022, ICIAM 2023, and 8ECM 2021
Initiator and spokesperson for major initiatives including
DFG Priority Program “Theoretical Foundations of Deep Learning”
Konrad Zuse School of Excellence in Reliable AI (relAI)
AI-Hub@LMU
With groundbreaking work across deep learning, harmonic analysis, inverse problems, imaging science, compressed sensing, and neuromorphic AI, she is one of the most influential voices shaping the future of Artificial Intelligence.
Trustworthy & Sustainable AI: Powering the Future with Next-Generation Computing | Gitta Kutyniok | Ludwig-Maximilians-Universität München
Artificial Intelligence is driving a new industrial revolution—transforming science, industry, communication, medicine, and every aspect of modern life. But as AI systems grow in scale and influence, two fundamental challenges stand in the way of a reliable, long-term future: trustworthiness and sustainability.
Follow SAI Conferences on Linkedin: linkedin.com/company/saiconference
Conference Website: saiconference.com/FTC
In this visionary keynote, Prof. Gitta Kutyniok explores how addressing these challenges requires going beyond algorithms and software—and taking a foundational look at the limits of current computing architectures. Through a deep, yet accessible, journey across mathematics, machine learning, hardware design, and neuromorphic computing, she explains why the next breakthroughs in AI may come from rethinking the very nature of computation.
🔍 What You’ll Learn in This Talk
1. Why Trustworthy AI Is Still an Unsolved Challenge
Safety, reliability, and security risks in modern AI systems
Black-box decisions, bias, privacy violations, and the global trust deficit
The mathematical foundations behind reliability—and what theoretical guarantees can (and cannot) offer
How generalization, expressivity, and explainability drive trustworthy AI
2. The Sustainability Crisis in AI
Why AI’s energy consumption is nearing global production limits
The hidden cost of training, communication, and large-scale inference
Insights from the U.S. decadal roadmap on semiconductor and AI energy consumption
Why simply scaling hardware is no longer viable
3. The Limits of Digital Computing
How traditional digital architectures contribute to reliability and energy challenges
What computability theory tells us about what neural networks can and cannot compute
Why certain inverse problems remain fundamentally non-computable on digital hardware
4. The Case for Next-Generation Computing
How analog, neuromorphic, and brain-inspired chips can overcome foundational barriers
Real examples of neuromorphic hardware (e.g., SpiNNaker) and its energy advantages
Spiking Neural Networks (SNNs): mathematical perspective, biological realism, and efficiency
Why time as a computational dimension opens entirely new AI capabilities
The promise (and limits) of emerging paradigms like optical computing, quantum computing, and bio-computing
5. Building a Future of Reliable & Green AI
Inside Bavaria’s state-level initiative Next Generation AI Computing
The importance of co-evolving hardware and software, rather than designing them separately
Applications in robotics, communication, medicine, and mission-critical systems
A vision for AI systems that are simultaneously powerful, trustworthy, and energy-efficient
🎤 About the Speaker: Prof. Gitta Kutyniok
Prof. Gitta Kutyniok holds the Bavarian AI Chair for Mathematical Foundations of Artificial Intelligence at the Ludwig-Maximilians-Universität München (LMU). She is affiliated with the DLR – German Aerospace Center and the University of Tromsø, and is a globally respected leader at the intersection of mathematics, AI, deep learning theory, and next-generation computing.
Her distinguished career includes:
Founder of EcoLogic Computing GmbH
Member of the Berlin-Brandenburg Academy of Sciences and Humanities and the European Academy of Sciences
Fellow of SIAM (2019) and IEEE (2024)
Plenary and invited lectures at ICM 2022, ICIAM 2023, and 8ECM 2021
Initiator and spokesperson for major initiatives including
DFG Priority Program “Theoretical Foundations of Deep Learning”
Konrad Zuse School of Excellence in Reliable AI (relAI)
AI-Hub@LMU
With groundbreaking work across deep learning, harmonic analysis, inverse problems, imaging science, compressed sensing, and neuromorphic AI, she is one of the most influential voices shaping the future of Artificial Intelligence.
