At the World Governments Summit (WGS), Robin Li, Co-Founder and CEO of Baidu, discussed the challenges posed by the rapid advancements in Artificial Intelligence (AI) and explored how emerging technologies are shaping the future of autonomous driving and cloud computing.

During a thought-provoking discussion with Omar Sultan AlOlama, UAE Minister of State for Artificial Intelligence, Digital Economy, and Remote Work Applications, and Vice Chair of WGS, Li stressed that innovation is unpredictable.

“Innovation cannot be planned. You just don’t know when and where innovation comes. What you can do is just foster an environment that is conducive to innovation.”

Highlighting the unprecedented speed of AI progress, he noted that costs associated with large language models (LLMs) have dropped by more than 90% within a year – far outpacing the historical trajectory of the computer revolution.

Baidu, with its deep roots in search engine technology, has leveraged this background to stay at the forefront of AI development. In March 2023, the company launched Ernie Bot, China’s first public ChatGPT-like AI application, just three months after OpenAI introduced ChatGPT. “It’s a very exciting time. We see innovations everywhere, and we have to adjust ourselves to these kinds of fast-changing innovations,” said Li.

Addressing the notion that China’s focus on reducing AI costs stems solely from geopolitical constraints on chip access, Li presented a broader historical perspective. He pointed to China’s approach to ride-hailing and autonomous driving as prime examples. Unlike Western markets, where higher fares make expensive autonomous vehicle technologies viable, China’s lower ride-hailing prices necessitate more cost-efficient solutions to bring down costs.

Discussing Baidu’s autonomous vehicle programme, Apollo Go, Li compared different technological approaches. While Tesla relies solely on cameras for its self-driving technology, Baidu employs a sensor-based system that enables fully driverless operations in designated areas from the outset.

Robotaxi services are already an integral part of daily life in China, demonstrating AI’s transformative potential in urban mobility, according to Li.

He added that the deployment of an autonomous driving service in a new city takes about two weeks, but due to regulatory requirements, the full implementation process typically requires six months, underscoring the challenge governments face in balancing innovation with necessary oversight.

When asked about AI’s risks, Li acknowledged concerns about regulatory readiness and rapid technological evolution. While Baidu’s Robotaxi service has proven significantly safer than human drivers, regulatory hurdles remain a key challenge.

Li also addressed the recent volatility in AI-related stock markets following the rise of open-source models like DeepSeek, which significantly lowered the cost of AI deployment. Despite concerns over diminishing returns on AI investments, he remained optimistic, encouraging key players to stay at the forefront of this revolution to keep pace with the rapid development of technology.

On the debate between open-source and closed-source AI models, Li observed that open-source models have gained traction by accelerating adoption and fostering broader innovation.

As AI continues to redefine industries worldwide, Li’s insights at WGS underscored the need for balance between innovation, cost efficiency, and regulatory readiness. With companies like Baidu at the helm, the future of AI promises to be both transformational and increasingly accessible to global markets.