The perfect result each time, there would be less scope for Google to make money by giving you results someone else might have paid for.
Which is what made the 12 minutes of I/O dedicated to search so significant. Very quickly, Cathy Edwards, a vice president for engineering, was showing that generative AI can help you buy stuff—in her example, a commuter bike. The prompt she demonstrated was “good bike for a 5-mile commute with hills.” The idiom might have sounded similar to the way we’d typically engage with Google, but it’s got a level of complexity with which we’d never typically challenge the search engine. The experimental product returned a few possibilities and potential follow-up questions.
Ostensibly, this was just another example of how AI, paired with Google’s vast data sets, can create a pretty cool customer experience. But Google has been trying to crack e-commerce for years: A sizable proportion of online shopping searches start at Amazon, completely circumventing Google. It’s been one of the company’s real weaknesses. Now it’s signaling that AI, where Amazon.com Inc. appears to lag Google technologically, offers it a chance to catch up. While the conversation focuses on Google’s battle with Microsoft, it heralds an impending fight with Amazon, too.
For all of AI’s amazing potential, the companies leading the charge technologically have the ability to set the parameters for how we engage with it, to dictate the nature of the conversation. And, as with the last generation of the web, it looks like they really want to use it to sell us more stuff.
