Google, Amazon, and Microsoft have a vision for generative AI. Will it work?
Sara Morrison is a senior Vox reporter who has covered data privacy, antitrust, and Big Tech’s power over us all for the site since 2019.
We got a trio of generative AI announcements from three Big Tech companies this week. Google said on Tuesday that it is extending Bard to several of its apps, including Gmail and Docs. The next day, Amazon revealed that it will let you have “near-human-like conversations” with Alexa “soon.” On Thursday, Microsoft held an event to announce that it plans to embed its generative AI assistant, “Copilot,” across many of its products.
The products and services are different, but the idea the companies behind them are selling is the same: Generative AI is amazing and our generative AI tools are amazing, so we’re going to embed them in as many of our services as possible to make your life amazing. Or, as Colette Stallbaumer, general manager of Microsoft 365 said at the company’s Thursday event: “Soon, you won’t be able to imagine your life without it.”
You just have to imagine your life with it first because it’s not here yet. And then you have to wonder if people will really use these tools when they do roll out. This isn’t the first time tech companies have gone big on intelligent assistants, only for the public to either hate them or be largely indifferent to them. We can go back to the late ’90s with Clippy, Microsoft’s notoriously much-loathed Office assistant. More recently, we’ve gotten smart assistants like Alexa, Apple’s Siri, Google’s Google Assistant, and Microsoft’s Cortana. It’s safe to say that those haven’t gotten the kind of adoption their makers hoped for, both in how many people are using them and how many things they’re using them for. Microsoft gave up on Cortana-powered smart speakers long ago and will soon stop supporting it just in time for its generative AI tools to take over. Amazon, on the other hand, is resting its hopes for Alexa on generative AI, which it calls its “north star.”
It doesn’t help tech companies’ case that these next-generation assistants we’re supposed to use for everything have already had some high-profile flubs. That makes it hard to trust both what these chatbots tell us and that they’ll be able to do what their developers claim anytime soon. Older digital assistants were far from perfect, but the stakes were a lot lower. There are real consequences when chatbots fall short. Alexa playing “Desperado” when you asked it to play “Despacito” is annoying. ChatGPT inserting a bunch of false information that it insists is correct into an important work document could get you (and potentially many others) in a lot of trouble.
Yet Microsoft continues to push especially hard on this vision of a generative AI-fueled personal assistant that knows you and helps you across your digital life (it also appears to be the furthest along in developing it, not to mention that $13 billion partnership with OpenAI, the hottest generative AI company out there right now). The company’s internet search announcement back in February was a big deal, and it likely spurred Google to roll out its internet search “experiment,” Bard, shortly afterward. If Microsoft hadn’t jumped first, Google may well have continued to take its time perfecting Bard before releasing it. Microsoft CEO Satya Nadella even took a swipe at Google at Thursday’s event, saying, “We are striving to breathe some innovation and life” into a “market that’s dominated by one player.” He didn’t call out Google by name, but the company is currently on trial over its dominance of the search market. And also: Duh.
After all that, most people still aren’t using Bing, which saw only a tiny bump in usage after it got its supposedly revolutionary chatbot. Bard doesn’t seem to have generated much interest either, although Google hasn’t played it up the way Microsoft has for Bing. Neither company releases much data on how many people are using these new products, but the stats we do have suggest it’s not very many. Interest in consumer-facing generative AI seems to have fallen off after the late 2022 and early 2023 burst of excitement when they first released.
Some of this is to be expected. Buzz and curiosity tend to be short-lived. But some of it is likely due to people not finding much use for AI chatbot internet search in their daily lives. The fact that ChatGPT usage went down in the summer and is going back up again in the fall indicates that a lot of its use is from students, who may well be using it to do their homework and write essays for them rather than learn information that allows them to do that homework and write those essays themselves.
So now companies have turned their attention toward using generative AI to assist us well beyond searching the internet, which was probably their goal all along anyway. Google and Microsoft do a great job of playing up their products’ strengths and how much better their AI tools will work once they’re in and working across everything, rather than siloed within each app or service. In theory, they’ll be able to combine their tremendous libraries of preexisting data and knowledge with the data and knowledge they have about their users, giving us humans a personalized, efficient assistant that vastly improves our work and personal lives. The possibilities are endless, everything works seamlessly, and we’ll all be able to focus our time and energy on more important things. There’s a reason Microsoft calls its assistant Copilot and uses words like “companion” when describing what it will do, and be, for you.
The reality is that we aren’t there yet. Some of Microsoft’s AI tools won’t be here until next year, and there is no date for when Microsoft’s ultimate vision, the ability to use Copilot across all of its products, will be realized. As for what we have now, you still have to check (or you really should check) everything chatbots do and tell you for accuracy. That will only be more important if and when people integrate them into really important parts of their lives. Google just rolled out a new tool so users can do just that. That’s helpful, but it’s also an admission — and Google continues to call Bard an “experiment,” further reinforcing the idea that this is something to try out but not fully rely on.
You can see the new wave of generative AI assistants as an example of progress, with tech companies making ever-better digital assistants that are sure to catch on once they’re good enough. You might also see it as tech companies trying to push something on people that they just don’t want or need in an effort to capture as many parts of their lives as possible. So far, it’s mostly been the latter. But maybe Copilot will be the Clippy Microsoft always knew we wanted.
Source: vox.com
