Google has also been running a parallel experiment with using AI to remake its main search interface. launching a generative search experience which offers chatbot-like responses before the familiar list of ads and links.
The company said just a few weeks ago that does not anticipate a “moment of light change””when the generative search experience completely replaces Google Search as we know it. But Google plans to push “the limits of what’s possible” and think about “what use cases are useful and that we have the right balance between latency, quality and feasibility,” Liz Reid, vice president and general manager of Search, said in the time. Like Pichai, she seems to think it’s time to experiment with some radical alternatives to Google’s established model.
Pichai says Google is focused right now on getting the generative AI experience right, but is “open to possibilities around” paid and ad-supported generative AI experiences. He declines to say whether Gemini’s paid offering will remain entirely ad-free, but he pointed to another Google-owned product where it’s possible to remove ads entirely.
“YouTube has been a very good example of this,” says Pichai, referring to the paid and ad-free level than YouTube I started experimenting several years ago. “Ads allow us to offer products to more people, but there will be cases of subscriptions that allow people to get a different experience.” He adds, “I can imagine the same user going back and forth between a free search and a Gemini subscription.” In other words, generative search would no longer be an accompaniment to search, but rather a main menu item, albeit a more expensive one.
Caution is advised
There’s another big reason why Google might want to charge money for its AI services: it helps defray the huge computing costs associated with training and running a large language model.
“We can make projections over our 25 years: If something on day zero costs this month, how much will it cost to do the same task a year from now, and so on?” Pichai says. “We’ve taken into account the efficiencies we’ll get in the underlying models and then priced them in a way that we think makes sense.”
Whatever Google’s motivations behind selling chatbot subscriptions, the technology it offers has to work reliably. Pichai acknowledges that Google Gemini, even the advanced version, still runs the risk of hallucinating, as bard did iteither as other generative AI applications have done. “We want people to be aware of it,” says Pichai. “I think technology is useful for many people. But it has to be used in the right way and I still worry about people trusting it.”
Pichai says, of course, that Google is trying to reduce the phenomenon of models run amok. But he also cautions that the word “hallucinate” should be used carefully and suggests that hallucinating was a feature as well as a bug, which is a fascinating twist on misinformation. He believes that technology should be based on real events, but if you boil it down too much, his chatbot becomes very boring very quickly.
A generative AI experience should be “imaginative,” says Pichai. “Like a child who doesn’t know what the limitations are when he imagines something.” Kind of like the early days of the web.