I’ll put up with the AI slop if Demis Hassabis actually saves the world

During the last few minutes of Google’s typically marathon-length Google I/O developer conference keynote, Deepmind co-founder Demis Hassabis said:

“I’ve always believed the number one application of AI should be to improve human health,” before adding, “With the goal of one day, solving all disease. Something that would have seemed impossible just a few years ago.”

A worthy AI end goal

I’m onboard with it

His words followed a short segment on how Google’s AI is being used in science and medicine, global simulations to measure food distribution and deforestation, and in prediction models to better understand extreme weather phenomena.

Not only was it inspirational, but it was also positive and human-centric, giving us a glimpse of how incredibly powerful tools are being used by very smart people to make the world a better place.

It’s AI I can get behind, and I doubt I’m the only one. If Hassabis’s vision truly is one of the end goals for AI, I may just about be able to put up with all the slop. Because what preceded Hassibis’s declaration was a lot of slop.

Troughs and troughs of slop

Search for, and buy your slop using AI

Google I/O was absolutely stuffed full of AI slop, to the point where understanding what the tools being shown were actually for, was impossible. Here are just a few examples.

Gemini Omni was demonstrated by creating AI videos created from a single source. Gemini Omni made otherworldly environments, sci-fi vistas, underwater worlds, and other unimaginable scenarios. Yes, it looked cool, but it was slop. Pure slop.

Google’s Search bar removes the need to actually look at websites and sources, and instead serves up slop, which, when you pay attention, still requires you to look through the sources, because it’s rarely 100% correct.

Google also showed how its AI agents will be able to help you shop for physical slop, driving a need to buy stuff in order to feel better about our empty existence, at the lowest possible prices because we’re so poor because AI has taken out jobs.

Slop on your watch, and in your glasses

A never ending stream

Even besides this, Google dedicated an entire section of the keynote to AI media generation, under the guise of “bringing ideas to life.” Google Pics takes images and turns them into something else, Stitch helps you design your own slop, and Google Flow can take your own videos, music, and photos and turn it into slop.

Google plastered its display with the phrase “AI for human creativity,” but it’s really not. It’s imagination-free technology creating something outlandish and unrealistic from a basic prompt. The slop overload made it hard to tell all the different tools apart, because they were mostly demonstrated using AI-generated images and video, leaving me to assume that’s all they can do.

It couldn’t even find a way to demonstrate Gemini on Samsung and Google’s smartglasses without it, preferring to take a photo of the crowd and make it a cartoon with a blimp flying overhead. The worst feature? It pushed the resulting slop image to a smartwatch. Just like there was no escape from all this during I/O, there’s going to be no escape from it on any of your devices either.

Big ambitions

Let the experts speak

Hassabis initially said AI may be able to help cure disease over the next decade or so in an interview in 2025. It sounds almost too incredible to be true, like something we’d dismiss as being so completely unlikely, the person telling us has to be deluded.

However, Hassabis is a Nobel-prize winning AI expert dedicated to advancing the technology, and committed to its use in science and medicine. If we’re going to be convinced by anyone, it’ll be by someone like Hassabis.

I was more engaged by Hassabis’s closing speech at I/O 2026 than anything in the 90 minutes before it. It fired my imagination, and made me excited to be a part of the generation where we could see such a meaningful change to everyone’s life through technological advancement.

Can we concentrate on disease, please?

Seems more important

I could have listened to Hassabis talk about how it may be possible, and what has already been achieved, for a lot longer. Instead, Google dedicated so much of its most important presentation of the year to showing us how AI can make even more convincing slop, even faster than before.

If using AI to cure disease, or coming close to it, means training it using endless image generation tools, yet more ways of summarizing things, or creating massive emails to send to people who will just use an AI-generated reply in response, then it’s a trade-off most would accept.

But you can’t bury such lofty ambitions in a turgid lake of rancid slop. The golf claps echoing around I/O’s amphitheater shouldn’t be taken as widespread acceptance of this nonsense. If you have to dumb AI down to explain it, Google, you can still use the backdrop of curing disease. It’ll be a lot more compelling than pretending yet another video where AI makes you look like an alien is entertaining, and a worthy use of a technology that apparently has the potential to cure all our ills.

To quote Marvin the robot from The Hitchhiker's Guide to the Galaxy, commenting on its underused intellect, "Here I am, brain the size of a planet, and they ask me to take you down to the bridge."