Going FAR with Foundation Models – Fair, Accurate, Robust AI for All Data
Unlock robust, AI-assisted decision-making that performs consistently – no matter the data type, customer base, or market conditions. From structured spreadsheets to unstructured text and images, Astute AI adapts and delivers.
About Us
Astute AI is a collaborative research and innovation project led by a team from the University of Sussex (UK) and the Basque Center for Applied Mathematics (Spain). With support from the European Research Council and the UKRI Higher Education Innovation Fund, we are developing a new generation of decision-making systems powered by foundation models that are fair, accurate, and robust.
Our team brings together leading researchers in machine learning, algorithmic fairness, and robustness—backed by decades of academic and commercial experience across AI and venture capital.
The Problem
Modern organisations face a flood of complex, inconsistent data – spreadsheets, text, images, sensors, and more. Yet most AI systems can only handle one type at a time.
This leads to:
- Underperforming models in real-world use
- Costly manual integration of new data
- Unreliable or biased outcomes, especially with changing customers or markets
For businesses making high-stakes decisions, these limitations carry serious risk.
The Solution
One system, multiple data types
No need to change your existing workflows
Robustness to change
Causal data simulation
