Page 21 - The Annual AI Governance Report 2025 Steering the Future of AI

The Annual AI Governance Report 2025: Steering the Future of AI

how electrons interact or how materials respond to light. Another major stream is autonomous
37
experimentation, where robotics and large language models (LLMs) are combined to design
and execute experiments dynamically. This creates a closed-loop system of AI-guided synthesis,
real-time feedback, and iterative refinement, dramatically shortening the development cycle for
new materials—from decades to potentially months. AI scientists are specialized scientific agents
that can reason, plan experiments, interpret data, and collaborate with human researchers in
iterative discovery processes. Their value extends to designing modular AI tools that can be
38
reused across disciplines, enabling scalability and adaptability in scientific workflows. 39

New Value Streams in Transportation
Emerging value streams associated with autonomous vehicles (AVs) span across multiple sectors
and are poised to reshape the transportation, logistics, urban planning, and public service
landscapes. In logistics and delivery services, AVs create opportunities to reduce labour costs
and enhance delivery speed, particularly in addressing the “last-mile” delivery challenge. For
instance, the integration of autonomous drones and robots can improve service efficiency
and reach underserved or hard-to-access locations. In the transport sector more broadly,
AVs can reduce operational costs through improved fuel efficiency, predictive maintenance,
and reduced accident rates—yielding potential economic benefits estimated at £51 billion
annually in the UK and up to $936 billion per year in the U.S. Additionally, AVs are expected to
improve personal mobility for individuals with limited access to transport—such as older adults
or those with disabilities—thus generating new demand for inclusive mobility services. Fleet-
based shared AVs may also disrupt traditional public transit models, offering dynamic, low-cost
alternatives that could be particularly effective in first- and last-mile connections. Moreover, AVs
can enable productivity gains by allowing passengers to work during travel and by reducing
traffic congestion, which in turn can boost regional economic integration and increase real
estate values in peripheral urban areas. 40

New Value Streams in Agriculture

New value streams in agriculture are emerging through the combined use of analytical and
generative AI, which together have the potential to unlock an estimated $250 billion in economic
value globally—$100 billion on-farm (“on the acre”) and $150 billion at the enterprise level.
On the farm, AI supports yield optimization through virtual agronomy advisors that integrate
weather, soil, and pest data, while also enabling labor efficiencies and input cost reductions
via precision agriculture. At the enterprise level, generative AI is driving innovation across
R&D, marketing, operations, and supply chain management—generating hypotheses for crop
development, personalizing customer outreach, and automating regulatory processes. 41

37 Chen, C. (2024). AI in materials science: Charting the course to Nobel-worthy breakthroughs. Matter, 7(12),
4123–4125.
38 Gridach, M., Nanavati, J., Abidine, K. Z. E., Mendes, L., & Mack, C. (2025, March 12). Agentic AI for Scientific
Discovery: A survey of progress, challenges, and future directions. arXiv.org.
39 Lu, C., Lu, C., Lange, R. T., Foerster, J., Clune, J., & Ha, D. (2024, August 12). The AI Scientist: Towards Fully
Automated Open-Ended Scientific Discovery. arXiv.org.
40 Thomas, F. (2024, December 17). What Might be the Economic Implications of Autonomous Vehicles? NIESR.
41 Nuscheler, D., Fiocco, D., Prabhala, P., Perdur, R. M., Brennan, T., & Gautam, Y. (2024). From bytes to bushels:
How gen AI can shape the future of agriculture. McKinsey & Company.

12