Google’s AI future is Dramatically Different from its Past

Google’s recently-concluded I/O developer conference didn’t try to hide the fact that it was an AI-filled buzzword fest, and while a lot of Google’s demos during the keynote presentation gave us a clear preview its AI prospects, some folks (myself included) couldn’t help but feel that we’re seeing history repeat itself. With that in mind, you might be asking “how so?” and that is a fair question, after all.

READ: AI was at the forefront of Google I/O, but so was Android

To answer it, all one has to do is simply look at Google’s recent AI-fuelled announcements and compare it with its past products – during I/O, we were introduced to several different AI models including Gemini Pro 1.5, Gemma 2, Gemini Flash, Project Astra, Imagen, Veo, and a whole lot more. Looking back, the hype behind these new announcements was somewhat akin to the launch of Google’s ill-fated Stadia platform, topped with a serving of Google’s somewhat messy history with multiple failed products, such as its many messaging apps.

That being said though, there are some key differences this time around, especially with regards to how these new announcements work for Google.

For one, it could be said that this new lineup of multiple AI models – besides being new ways for Google to show off its AI development – are primarily revenue generators for the company. Despite its Pixel smartphone line growing in terms of sales, it’s clear that Google sees AI as a better way to go about earning profits (hence the absence of any Pixel news at I/O 2024). One could even say that Google’s dedicated focus on AI functions a business to business product, with leftovers being rolled out retooled for consumers.

One could even say that Google’s dedicated focus on AI functions a business to business product, with leftovers being rolled out retooled for consumers.

When viewed this way, Google might be aiming to fundamentally changing its business model from making money via ads to making money from selling AI computation and such.

Another difference is that the various AI models that Google showed off (at least at the moment) are designed to do different tasks, as opposed to multiple apps designed to do essentially a single task (i.e. Allo, Hangouts, Messenger, Voice, etc). There’s less repetition, despite the appearance of an overwhelming number of AI models.

Of course, there’s still the issue of commitment, at least on Google’s behalf – the Mountain View tech giant is infamous for undertaking multiple new projects at a time, only to end up shutting those it deems unsuccessful enough. At this point in time many are still doubting the longevity of AI, especially with Google at the helm.

We go back though to the argument of a potentially huge revenue stream for Google – if it plays its cards right, AI could end up being a massive win for the company, which in theory shouldn’t be a problem given its vast resources in terms of technology and funding. In any case, there’s a lot of caution coming from both onlookers and investors, and it should be a very interesting couple of years ahead.