While many people might see this as an anti-Google blog, I don't at all. And in that spirit I want to call out some great things that Google announced at Google I/O this week.
While I agree with the cynics that this is a form of privacy theater, it really marks a turning point in the market where Google feels enough pressure by consumer demand to, at the very least, pretend to take privacy seriously.
Auto-Delete Data: Users will be able to set the timeframe that Google can keep and use their data. After that time, it will be deleted automatically.
Incognito Across Apps: An incognito mode will be coming to Maps and YouTube, presumably meaning that what you do while in Incognito is not shared with Google.
AI on Devices: Some voice and facial recognition will now be done on device, which means your conversations and video don't necessarily need to live on servers to offer AI features.
Machine Learning at the Edge: Not only can you use voice recognition models without sending your data to a server, in some cases you'll be able to train those models on device as well, so Google only gets the improved model and not your personal data that trained it.
Ad Transparency: A new browser extension will allow you to see the chain of entities behind an ad that you are seeing.
The Mainstream Milestone
These product changes will theoretically make it harder for Google to target you with advertising. The less data they have on you, the less valuable the ads they can sell to advertisers.
Yet at the same time we see that roughly $30 billion of its $36 billion Q1 revenue was advertising.
When was the last time you saw a company pro-actively hamstring its business to meet users' demands? While I would guess that they've found ways to avoid these features eating into their bottom line, to me it signals that privacy is an even bigger deal to consumers than it would seem based on media coverage. It is officially a mainstream must-have.
Many of the coolest things Google wants to do are contingent upon most of your data being on their services. Duplex, the company's AI assistant, can only seamlessly book restaurants, flights, and hotels for you if it has access to your emails and calendar. If you're anonymous on the platform, their techno-utopia is not possible.
So giving more privacy controls to keep people on the platforms makes sense. And we'll only see a bigger move in this direction as consumers demand more protections. But it does seem like Google will need to make money through something other than advertising if it wants to keep up with users' evolving expectations. Otherwise the strategy will always be to give up the bare minimum of control to keep people happy.
What About Data Ownership?
My issue with all the main tech companies is that they are too big and entrenched to see the larger picture, which is not privacy, but data ownership.
It's great that more AI models will be trained on device instead of a less-secure server, but the value that my personal data is providing is still going to Google for free*. The company gets a better voice recognition model even if they don't get my voice recordings. Ideally, I would own and control my recordings and I could license them to Google and any other company interested in the training data.
Our faces, movements, and voices are training up the next generation of AIs. What we're offered in return is a vague promise that these AIs will benefit us somehow. Instead, we should be demanding real cash for our contributions and clear licensing that says how our data can be used, and for how long.
The market competition should be to produce the best products that bring value to a user's life. We can't get there without controlling and owning all our data ourselves. When we don't, we get our current status quo: Companies optimizing how to capture as much data from a user while keeping that user minimally happy.
* It's true that a user will benefit from Google's products getting better, but it's a very watered down benefit compared to the billion dollar businesses that Google is able to build off of our data in aggregate.