Google and Tel Aviv University Launch Three-Year AI Research Program
Google Israel and Tel Aviv University (TAU) have launched a three-year program to advance foundational research in artificial intelligence and data science. The work centers on language models, privacy-preserving AI, and algorithmic efficiency-areas that matter for scaling real systems and testing theory against practice.
Program Scope
The initiative will prioritize core research problems: innovation in language model design and training, methods that protect privacy by default, and techniques that cut compute and sample costs. Expect collaborations that connect computer science, engineering, and data-centric research with use cases in the life sciences, humanities, and social sciences.
Leadership and Funding
The program is led by TAU's Center for AI and Data Science, headed by Professor Yishay Mansour of the Blavatnik School of Computer Science and AI. Google's philanthropic arm is providing $1 million in support to fund research, scholarships, and community-building efforts.
TAU President Professor Ariel Porat noted the partnership builds on about five years of joint work spanning multiple disciplines. The new phase includes scholarships for PhD students and funding for students from Israel's periphery to widen access to advanced study.
Research Priorities
Professor Mansour highlighted a central challenge: building theories that explain why current AI methods work as well as they do. By probing how large language models are trained, the team expects to open new directions and improve the efficiency of learning processes.
Yossi Matias, Google's VP of Engineering and Research and head of Google Research, emphasized that academic excellence and industry collaboration are essential-cross-pollination between diverse disciplines often produces meaningful advances.
Inclusion and Community
Over the past three years, Google has also supported TAU's ExactShe program, led by Professor Tova Milo, dean of the Faculty of Exact Sciences. ExactShe focuses on building a supportive community for women in research-a foundation that strengthens future AI scholarship and leadership.
Why This Matters for Researchers
- Language models: opportunities to stress-test new training signals, data curation strategies, and evaluation protocols.
- Privacy: methods that reduce exposure risk while preserving utility-differential privacy variants, secure learning setups, and post-training safeguards.
- Efficiency: algorithmic improvements that lower compute, memory, and data requirements without compromising quality.
What to Watch Next
- Calls for PhD scholarships and cross-lab collaborations.
- Publications on theory of training dynamics for large models.
- Benchmarks, datasets, or toolkits that operationalize privacy and efficiency.
- Workshops or seminars hosted by TAU's Center for AI and Data Science.
Learn more at Google Research and Tel Aviv University.
If you're skilling up a research team, see curated options by role at Complete AI Training.
Your membership also unlocks:
