Podcast Summary: "The End of Human Driving? with Uber CEO Dara Khosrowshahi"

On with Kara Swisher, December 18, 2025

Episode Overview

This live episode of On with Kara Swisher, taped at Johns Hopkins University, features an in-depth conversation with Uber CEO Dara Khosrowshahi. The main theme revolves around the rise of artificial intelligence (AI) in business, particularly its application within Uber, and the future of autonomous vehicles (AVs). Swisher and Khosrowshahi explore the technological, regulatory, ethical, and societal implications of replacing human drivers with AI, how Uber is positioning itself in this shift, and what the future of mobility could look like.

Key Discussion Points & Insights

1. Uber as an Applied AI Company

• Uber’s DNA: Uber considers itself an applied AI company, where machine learning is foundational to most aspects of its operation—pricing, ride matching, route optimization, Eats recommendations, fraud detection, and even verifying drivers’ identities. (03:42–05:40)
  • Quote: “AI has been part of our genetics for a very, very long time now... We’re a technology company that operates in the real world.” — Dara Khosrowshahi [03:42]
• Data is Core: Uber collects vast amounts of data on user preferences and behavioral patterns, which fuels model improvements. (05:40–06:07)
• AI’s Immediate Impact: The biggest gains so far are in developer productivity—AI agents now monitor systems, debug issues, and suggest solutions faster than humans alone. (07:25–08:40)
  • Quote: “80, 90% of our developers... are using AI developer tools... Now we have AI agents that are on call." — Dara Khosrowshahi [07:25]

2. Where AI Works (and Where It Doesn’t)

• Success in Internal Tools: AI is effective in coding, documentation, and operational monitoring.
• Challenges in Customer Service: Attempts to automate customer service interactions with AI initially created double work as agents distrusted AI-generated conclusions, sometimes due to hallucinations. Uber is now trialing end-to-end AI solutions for low-risk scenarios. (09:57–13:12)
  • Quote: “We are having [AI] solve problems where the cost of a mistake isn’t very high... and then we'll learn from that.” — Dara Khosrowshahi [12:43]

3. Autonomous Vehicles—Promise and Reality

• Consumer Experience: Most first-time riders are amazed, but quickly adapt; AVs are described as "delightful" and offer unique benefits (privacy, reliability). (14:03–16:07)
• AVs on Uber’s Platform: Uber’s approach is to partner with as many AV companies as possible, acting as a marketplace rather than a manufacturer. Current AV rides are a small fraction of Uber’s total, but expected to grow. (14:03–19:51)
  • Quote: “I want every great qualified driver on the platform. I want every great qualified robot driver on our platform as well.” — Dara Khosrowshahi [15:07]
• Safety and Scaling: AVs are expected to be much safer than human drivers, but mass deployment will take 10–20+ years. (16:49–19:51)

4. Business Model Evolution: From Drivers to Robot Fleets

• Ownership of Fleets: Uber does not plan to own the autonomous fleets. Instead, other companies (private equity, small entrepreneurs, institutional investors) will own and manage AV fleets, similar to how hotels are owned by REITs but branded by chains. (22:58–24:57)
  • Quote: “Marriott doesn't own any of its own hotels... you're going to have fleets... [and] Uber brings demand to these.” — Dara Khosrowshahi [24:13]
• Competition & Coopetition: Uber expects AV makers like Waymo and Tesla to run both direct-to-consumer channels and collaborate with platforms like Uber for greater utilization. (25:20–26:00)

5. Regulatory and Societal Implications

• Expert Question from David Plouffe [26:14]: Plouffe asks how city leaders should rethink urban design and public transit in light of ride sharing and AVs.
  • Accessibility: Ensuring AVs serve entire cities, not just wealthy, dense areas.
  • Congestion: Efficient matching is crucial to avoid empty vehicles creating more traffic.
  • Transition: Hybrid networks of human and AV drivers can ensure gradual, responsible change. (27:24–31:19)
• Labor Displacement: Uber admits it does not yet have a solution for inevitable job losses as AVs scale, but emphasizes truthful communication and a gradual transition, including offering alternative work (like AI data labeling). (40:16–43:18)
  • Quote: “We just don’t have an answer right now. And there will be, we have time, but we don’t have all the time in the world.” — Dara Khosrowshahi [44:42]

6. Environmental Goals, EVs, and AVs

• EV Challenges: Uber (and industry at large) lags behind its own EV adoption pledges, with North American EVs at just 9%. Khosrowshahi notes the transition is happening faster in Europe, and that AVs will all be EVs, which may help accelerate transition. (32:30–33:50)
  • Consumer Reluctance: Riders aren’t willing to pay more for EV rides, though some will accept longer waits.
• Joking About ETA Accuracy: Swisher jokes that Uber’s time estimates are “bullshit,” to which Khosrowshahi gamely parries. (33:50–34:13)
  • Quote: “You just notice when we get it wrong. You only notice when you get it wrong.” — Dara Khosrowshahi [33:55]

7. Delivery Robots, Drones, and the Future of 'Eats'

• Uber Eats Growth: Now approaching half of Uber’s business.
• Delivery Innovation: Experimenting with sidewalk robots, bike-lane AVs, and drones. Each has its limits (e.g., sidewalk robots for short trips). (34:49–37:08)
  • Quote: “I think our Eats business is ultimately potentially going to be bigger than our mobility business. Absolutely.” — Dara Khosrowshahi [36:58]

8. Liability and Safety

• Legal Responsibility: The operator of the software “driver” is legally responsible for accidents. Nonetheless, Uber employs strict internal safety standards for AVs on its platform. (37:08–39:26)
  • Societal Tolerance: Even a tenfold reduction in fatalities might not be “acceptable” to society if robots are to blame.
  • Quote: “Society expects humans to make mistakes, and when a robot makes a mistake, it’s unacceptable.” — Dara Khosrowshahi [38:55]

9. Global Competition: China's Rapid Progress

• China LEADS in Manufacturing: Chinese companies are far ahead at building affordable, AV-ready vehicles. U.S. still leads in software, but China’s manufacturing edge is widening. (43:18–44:21)

Notable Quotes & Memorable Moments

• “So if an AV is provably 50 times safer than a human being, do you think we should allow human beings to drive?”
  — Dara Khosrowshahi, [00:00] and [39:50]

• “We’re a technology company that operates in the real world.”
  — Dara Khosrowshahi, [03:42]

• “AI drivers are going to get better. They have been... this is going to be [a] safer method of transportation over a long period of time.”
  — Dara Khosrowshahi, [16:49]

• “Most people who experience AVs love it ... for the first five minutes they’re amazed, and then seven minutes they’re texting and it’s just like being in any car.”
  — Dara Khosrowshahi, [46:05]

• “You were the one who told me we’d never get profitable. We proved you wrong on that one."
  — Dara Khosrowshahi [34:09]

• “Our Eats business is ultimately potentially going to be bigger than our mobility business.”
  — Dara Khosrowshahi, [36:58]

• On customer wait times:
  Kara: “Those maps are bullshit.” [34:01]
  Dara: “They're not bullshit, Kara.” [34:05]

Important Timestamps

• 00:00 — Show opens with the key question: Should humans be allowed to drive if AVs are 50x safer?
• 03:42–05:40 — Dara describes Uber’s “applied AI” DNA.
• 07:25–08:40 — Biggest impact: AI-driven developer productivity and system monitoring.
• 09:57–13:12 — Challenges when automating customer service with AI.
• 14:03–16:07 — Consumer experience and safety focus of AVs.
• 16:49–19:51 — Scaling AVs and the long transition timeline.
• 22:58–24:57 — Who will own AV fleets in the future.
• 26:14–31:19 — David Plouffe’s expert question and discussion on policy, accessibility, and congestion.
• 32:30–33:50 — Discussion of Uber’s environmental goals and EV challenges.
• 34:49–37:08 — Uber Eats growth and delivery bot/drone innovation.
• 37:08–39:26 — AV liability and societal perceptions of safety.
• 43:18–44:21 — China’s manufacturing edge in EVs and AVs.
• 44:42–45:45 — What we’re getting right/wrong about the AV transition.
• 46:04–46:32 — End with humor and letting consumers choose between human and robot drivers.

Final Thoughts and Takeaways

• Gradual Transition: AVs will not fully replace human drivers anytime soon. The hybrid model will persist for decades.
• AI’s Pervasiveness: AI touches every dimension of Uber’s business, but integrating it often lacks instant ROI and requires continuous adaptation.
• Societal Questions Ahead: Safety, labor displacement, regulatory adaptation, and equitable access are complex issues with no simple fixes.
• Competitive Landscape: Uber aims to be a neutral, global platform — working with many AV manufacturers, not building its own.
• Consumer Experience: Most are won over by AVs after trying them, but choice will remain central for years to come.

“Most people who experience AVs love it … for the first five minutes they’re amazed, and then seven minutes they’re texting and it’s just like being in any car.”
— Dara Khosrowshahi [46:05]