Uber's decision to put its money where its former CEO now stands is a quiet admission that the future of mobility was never really about cars. It's about data. The company's renewed bet on Dara Khosrowshahi, the same executive who steered it through the messy business of mapping, autonomy, and the slow bleed of ride-hailing margins, signals something practical: the next competitive edge isn't a faster app or a sleeker vehicle. It's the ability to turn movement into intelligence. And if that sounds familiar, it's because the same logic is already reshaping how we think about spreadsheets, as we noted in Talking to My AI Clone Taught Me to Question the Tech. There, the question was whether an AI that mimics us actually understands us. Here, the question is whether an AI that predicts our routes can truly serve us, or just nudge us toward the most profitable path.
Let's be direct: we don't need another announcement about AI's potential. We need honest assessments of what these tools actually change. Uber's move is interesting not because it's bold, but because it's defensive. The company has spent years fighting regulators, drivers, and the physics of urban congestion. Its moat was never technology; it was scale. Now, with competitors like Waymo and Tesla pushing autonomous fleets, scale alone won't cut it. So Uber is going back to a familiar face, betting that Khosrowshahi's operational discipline can turn a ride-hailing giant into a logistics brain. That might work. But it also raises a question we've been circling in our own reporting: when AI starts making decisions for us, how much do we actually understand the trade-offs? In Verify Your AI's Understanding: A Simple Check for Tax Season, we wrote about the importance of testing whether an AI truly grasps the task at hand. The same applies here. Uber can claim its AI optimizes for efficiency, but if that optimization squeezes drivers or raises prices during surges, the user is left holding the cost.
This is less about Uber's leadership and more about a broader shift in how we should evaluate AI-powered services. We've seen the pattern before. A company slaps "AI" on a product, and suddenly every spreadsheet, every route, every recommendation is treated as if it were inevitable. But as we argued in Navigating AI/ML Job Requirements: A Shift in Expected Skills, the real challenge isn't building the model; it's understanding what the model is actually doing. For Uber, the risk isn't that the AI fails. It's that it works too well, optimizing for metrics that don't align with what riders or drivers actually need. If you're a user, don't ask whether Uber's AI is smart. Ask what it's optimizing for. That's the question that matters, and it's the one we'd push back on if a reader asked us directly.
The practical consequence is simple: watch what Uber does with its data, not what it says about AI. If Khosrowshahi's return leads to more transparent pricing models or clearer driver incentives, that's progress. If it just means more aggressive surge pricing, then the AI is working against you, not for you. The detail to watch is the next earnings call, specifically how much Uber spends on AI infrastructure versus driver acquisition. That number will tell you more than any press release. Because in the end, the spreadsheet analogy holds: a tool that automates your calculations is useful, but only if you understand the formulas underneath. Uber's formula is about to get more complex. Make sure you're reading the cells, not just the summary.