Hyderabad: Artificial intelligence can do a lot today, from doing research and writing essays to generating images, audio, and videos. While most of it stays in the digital world, AI has also found multiple use cases in the real world. One of the most intriguing of these is its role in autonomous taxi driving, with Waymo leading the charge as it reaches a milestone of 10 million rides.
Owned by Google-parent Alphabet, Waymo began in 2009 as the Google Self-Driving Car Project, with an aim to develop fully autonomous vehicles using AI, sensors, and mapping technology. They tested modified vehicles, equipped with LiDAR, radar, and cameras, and clocked 300,000 autonomous miles by 2012. The project became Waymo (a new way forward in mobility) in 2016.
Navigating through regulatory hurdles, public scepticism, legal challenges, and scrutiny, Waymo progressed from being an experimental research to a commercial leader in autonomous mobility. Starting with achieving a milestone of delivering the first fully driverless ride on public roads in Austin, Texas, the company followed through with initiating public trials in Phoenix, Arizona—offering free rides to early riders, scaling its fleet with partnerships, launching Waymo One—its first commercial self-driving ride-hailing service in Phoenix, launching its fully driverless rides without the presence of a human safety driver, expanding the service to cities like San Francisco and Los Angeles, and testing autonomous trucks for freight transport in Texas.
According to the data published by the California Public Utilities Commission and accessed by WSJ, Waymo was doing 10,000 paid driverless trips a week in August 2023, followed by up to 50,000 rides in May 2024. In August 2024, it reportedly hit 100,000 rides, which has now escalated to 250,000.
At the Google I/O 2025, Waymo co-CEO Dmitri Dolgov announced that the company reached 10 million autonomous rides in 2024 as it operated a fleet exceeding 700 vehicles and clocked over 50,000 weekly public rides across multiple cities—a figure which is five times higher than the same time last year.
Speaking at Google’s annual developer conference, Dolgov recalled how Google's founders, Larry Page and Sergey Brin, gave them two challenges in the beginning: drive 100,000 miles in autonomous mode and drive 10 routes. Dolgov proudly revealed that a team of about a dozen engineers completed both challenges in 18 months, using technology from 2009.
“However, there's a massive difference between doing 10 routes once or building a system that just works most of the time versus building a reliable, scalable, fully autonomous service with no human behind the wheel,” Dolgov said. “And the challenge isn't just building the tech, it's proving that it's safe. It's earning the public's trust. It's building a product that people love and actually want to use.”
At the heart of Waymo is cutting-edge technology, which is brought together by artificial intelligence. At Google’s annual developer conference this year, Dolgov discussed the critical role of AI in autonomous vehicles and shared how “Waymo represents the most mature application of AI in the physical world” today.
Kids chasing dogs, chasing balls on the streets of LA… once again, @Waymo AI with advanced sensing making our roads safer. pic.twitter.com/QGZxDtuKaU
— Dmitri Dolgov (@dmitri_dolgov) June 4, 2025
He outlined how Waymo's autonomous driving technology has evolved through the integration of increasingly powerful AI and machine learning models, from early breakthroughs with convolutional neural networks to the transformative role of transformers, large language models (LLMs), and visual language models (VLMs). These advancements have significantly enhanced the Waymo Driver’s perception, understanding of intent, decision-making, and trajectory prediction, Dolgov said, adding that VLMs in particular allow the system to interpret complex visual-text information, such as confusing street signs or dynamic driving scenes, with greater contextual awareness.
“With every mile driven, we're also making our roads safer," he said, referencing an analysis by insurance company Swiss Re, which used insurance claims as a proxy for fault to find that Waymo’s autonomous driver is about 10 times better than humans.
“The data speaks for itself," Dolgov said. “When it comes to serious collisions, anything involving an injury or airbag deployments, the Waymo driver is about five times safer than humans. And for anything involving pedestrians, the results are even stronger.”
With over 50 million fully autonomous miles under Waymo’s belt, Dolgov took the safety record as a badge of honour and a step forward toward the company’s mission of making roads safer. He explained the three key challenges of autonomous driving:
- Self-driving vehicles must navigate a dynamic and unpredictable physical environment, dealing with various weather conditions and human drivers who can behave erratically.
- The stakes are incredibly high, as any mistake can lead to severe consequences, so the system must operate with an extremely low error tolerance
- Autonomous driving occurs in real time, requiring immediate decision-making within milliseconds without the luxury of delayed data processing in a remote server.
Dolgov also elaborated on how AI enables the Waymo Driver to behave in a socially intelligent manner, not just detecting objects but predicting behaviours and responding proactively. For example, it can identify and interpret sparse cues—like a pedestrian’s feet moving beneath a bus—to take defensive actions well before a human driver might react.
He also highlighted the impressive responsiveness of the Waymo autonomous driver, showing examples from Austin. In one case, a scooter rider lost balance and fell onto the road, prompting the Waymo driver to swerve left to avoid a collision. In another instance, an oncoming truck unexpectedly moved into the Waymo driver's lane, leading the AI to instinctively shift onto the shoulder and prevent a head-on crash.
“Today, Waymo is operating a 24/7, fully autonomous ride-hailing service with no human behind the wheel in four major US markets, San Francisco, Phoenix, LA, and Austin,” Dolgov said, adding that the company will soon bring it to Atlanta, Miami, DC, and many more cities.
