Globally, the transportation sector accounts for about 23% of global greenhouse gas emissions, 42% of which comes from freight logistics alone. Could artificial intelligence (AI) help the freight sector operate more efficiently and therefore reduce these emissions?

This blog post will discuss the potential for AI to draw from the industry’s large volume of actionable data to meet climate goals, highlighting three use cases: route optimization, dwell time, and capacity optimization. Some of these applications are already in motion: as of 2024, major delivery companies like Amazon and the United States Postal Service have begun implementing AI in their shipping and logistics, producing millions of dollars in operational savings and emissions reductions. Now we’re seeing the emergence of several smaller startups that assist companies that cannot invest in their own technology, allowing anyone to capitalize on the AI revolution. 

Use Case #1: Route Optimization

A recent report from the World Economic Forum (WEF) estimated that companies can achieve a 7% reduction in transportation emissions by using AI to optimize routes. Using live variables and historical route data, AI can both create the most efficient long-term path for vehicles and adapt to real-time conditions, producing day-to-day fuel savings. WEF suggested that deploying route optimization tools across global road freight transport could achieve emissions reductions equivalent to taking approximately 25% of all heavy- and medium-duty trucks in the U.S. off the road.

Use Case #2: Dwell Time

A similar way companies can optimize fuel consumption is by decreasing vehicle dwell time. Dwell time is the amount of time a vehicle spends stationary, whether that is for loading/unloading freight or simply stopped in traffic. As of 2023, dwell time inefficiencies, primarily congestion, cost the U.S. trucking industry approximately $94.6 billion in lost productivity. An example of this AI implementation at scale is Hangzhou’s “City Brain” (HCB), which uses real-time data from various sources, such as traffic lights, cameras, and vehicle GPS systems, to monitor and manage traffic flow. The system can adjust traffic light timings to reduce congestion, identify optimal routes for emergency vehicles, and even detect traffic violations. After the HCB’s implementation, Hangzhou dropped from China’s 5th-most-congested city to its 57th, reducing total greenhouse gas emissions by 1.5%.

Use Case #3: Capacity Optimization

Another way AI can improve operational efficiency and reduce emissions is by optimizing load capacity. If the U.S. can reduce unnecessary trips, it can reduce freight emissions by up to 4%. Some studies show that empty miles constitute 20-35% of U.S. trucking miles, and the trucking industry loses over $150 billion annually due to empty capacity. Of the miles that are not empty, trucks are carrying an average of 57% of their maximum load. Using AI, industry and policy leaders can promote co-loading among freight companies, reduce empty miles through load matching, and optimize trailer packing configurations to improve freight loading efficiency. Companies like Flock Freight and Convoy have begun offering load pooling, with Flock Freight reporting cost savings of up to 20% and a net 51% reduction in miles traveled. Expanding this technology across the U.S. alone could reduce annual CO₂ emissions by 19 million tons. 

Moving Forward

While larger infrastructure or technology shifts are important and necessary long-term goals, they require heavy initial investments. Industries can be hesitant to commit to these investments due to their unfamiliarity with new technologies. Using AI, however, companies can achieve unprecedented efficiency gains and cut operational expenses, creating win-win outcomes: they can reduce climate impact while simultaneously improving operational efficiency. 

To realize these benefits responsibly, AI deployment must be governed by strong ethical frameworks that safeguard against social inequities, labor displacement, and new environmental risks. Meeting climate pledges will require continued innovation, rapid deployment of the best available AI tools, and policies that ensure these technologies are accessible so that organizations of all sizes can share in both the economic and environmental gains.

 

Authored by Eli Walchli, Student at UC Berkeley and Intern at the Center for Law, Energy and the Environment (CLEE).