At the 2025 AI for Good Global Summit, Xi Cao, Director of AI Standardization at the Artificial Intelligence Center, China Mobile Research Institute, presented a practical case study on how AI and advanced communication technologies are being used to support the conservation of the Chinese white dolphin. The talk outlined a collaborative initiative between China Mobile and eight partners, focused on deploying AI and 5G-A ISAC (Integrated Sensing and Communication) to promote a more harmonious coexistence between humans and marine wildlife.
A marine species under pressure
Cao began by introducing a rare and iconic species: the Chinese white dolphin. He explained that this marine mammal is an important indicator species for the health of marine ecosystems. Xiamen, he noted, is the only city in China where the Chinese white dolphin can still be observed within an urban setting.
The area is protected under the Xiamen Rare Marine Species National Nature Reserve, which shelters 12 rare species including dolphins, wild herons, and small but ecologically significant fish. However, Cao emphasized that human activities are increasingly encroaching on the dolphin’s habitat. The waters of Xiamen Bay have become heavily trafficked by ferries, fishing boats, engineering vessels, yachts, and speedboats.
Cao noted that continuous monitoring of the dolphin population revealed a negative impact due to these human disturbances.
“From 2004 to 2019, continuous observation of the Chinese white dolphin population indicated that, with increasing interference by human activities […], wildlife has been negatively impacted,” Cao explained.
He reported that 14 deaths were recorded during this period, several of which were directly linked to speeding vessels or illegal entry into protected areas.
Building a digital conservation solution
In response to these threats, China Mobile and its partners designed and implemented a comprehensive solution called “AI plus 5G-A enabled marine wildlife conservation.” The project, deployed in the Xiamen Reserve, integrates autonomous networking with both sensing and management technologies to monitor and protect the marine environment more effectively.
On the sensing side, the system leverages 5G-A ISAC, smart disaster response, dynamic inspection, and interference suppression tools. On the management side, it incorporates a Conservation Command Cockpit, intelligent planning features, and a mobile law enforcement solution.
According to Cao, traditional marine monitoring methods in use before this project, such as radar, electro-optical sensors, and the Automatic Identification System (AIS), suffered from high costs and limited efficiency. These older systems also had constraints such as limited range, poor performance in bad weather, and a heavy reliance on AIS equipment.
Integrating sensing and communication
To overcome those limitations, the team introduced a more advanced system combining 5G base stations, radar systems, virtual terminals, and standardized data sharing interfaces. This integration enables real-time data exchange and intelligent processing between vessels and onshore operations within the reserve.
However, Cao acknowledged that placing 5G base stations along the coastline introduced new challenges, especially due to exposure to harsh marine weather. Base station antennas could shift position or orientation, reducing their coverage and effectiveness.
To address this, China Mobile deployed a Smart Disaster Rescue Solution that uses AI to automatically identify abnormal antenna conditions. This improvement raised fault identification accuracy to 90 percent, cut recovery time by 70 percent, and boosted self-healing rates by 30 percent.
The system also adapts to the predictable patterns of vessel movement, both spatially and temporally. Using two advanced algorithms, a dynamic inspection system monitors real-time changes in vessel distribution and adjusts the antenna coverage accordingly. This allows the network to better coordinate communication and sensing services. As a result, coverage expanded by 10 percent and vessel detection accuracy increased from 80 percent to 97 percent.
Watch the full session here:
Suppressing interference and enhancing decision-making
Another technical advancement was the introduction of a dedicated signal pontoon in the interference suppression module. This component ensures that communication operations do not interfere with sensing functions, maintaining both low missed detection rates and low false alarm rates.
Cao then shifted to discussing the data management challenges that accompanied the sensing improvements. He pointed out that the diversity and volume of incoming data, along with limited capacity for comprehensive analysis, previously hindered effective decision-making.
To resolve this, the team created the Conservation Command Cockpit, which enables simultaneous analysis and visualization of multiple data streams. The cockpit automatically recommends rescue responses and enhances situational awareness for law enforcement and reserve staff. Cao highlighted that this system has resulted in a marked improvement in on-site analysis capabilities.
The project also addressed limitations in understanding dolphin behavior due to previously narrow data dimensions. A new monitoring model has been constructed that builds a multidimensional profile of Chinese white dolphin behavior and spatial patterns. This helps to assess migration and activity in greater detail, supporting more efficient protection and enforcement strategies.
Empowering frontline enforcement
Cao explained that the new system also reduces the strain on human resources. Previously, effective surveillance required two officers: one for field operations and one for back-office monitoring. The back-office officer had to remain on duty around the clock to relay data in real time. This created high labor costs and risks of delayed responses.
To counter this, the team developed a Mobile Law Enforcement Assistant solution. This mobile platform integrates real-time and historical data and enables on-site access through a digital interface. Officers can now perform enforcement tasks directly in the field. According to Cao, this innovation improved enforcement efficiency by 65 percent.
Measurable impact on conservation efforts
Cao summarized the project’s outcomes by emphasizing its successful commercial deployment and real-world validation. The system now provides continuous, round-the-clock monitoring of vessel movements. Real-time alerts prevent unauthorized entries and support rapid responses to emerging threats.
The conservation outcomes are significant. The solution has resulted in a 97 percent reduction in dolphin deaths caused by vessel collisions. The core dolphin habitat expanded by 60 percent. The area covered by real-time monitoring nearly tripled, increasing from 120 to 330 square kilometers. More than one million vessels are now tracked annually, and monitoring efficiency has improved by 65 percent.
Moreover, law enforcement patrols have been reduced by 50 percent, giving officers more time to rest and improving working conditions. Cao noted that this improvement enables safer and more sustainable field operations.
A shared responsibility
Cao concluded the session by highlighting the importance of continued awareness and cooperation in protecting marine species like the Chinese white dolphin. He expressed hope that the audience would come to appreciate the species as much as he does and extended an invitation to visit China to witness these dolphins in their natural habitat.
