Earlier this month, a Google Chrome vice president put out a blog post on the company’s website announcing three changes coming to Google Chrome, the most popular web browser in the world.
The first is a new tab organizer for those of us who like to keep dozens of web pages open as we work on projects. The second is a new feature that will allow folks to create custom wallpapers and themes for when they’re browsing the web.
Herrman joined Texas Standard to talk about what to expect. Listen to the interview above or read the transcript below.
This transcript has been edited lightly for clarity:
Texas Standard: Let’s talk about this third feature. AI is involved here, right?
John Herrman: Yeah. These days it feels like everything’s got a little bit of AI, but this is kind of the big one.
This is a feature called “Help me write” and Google is testing this currently and is planning on rolling it out to more users in the coming months. The basic idea is anywhere there’s a text box on the internet, Google Chrome will offer to give you a hand with filling it out. That might be a product review, a restaurant review, text box for a social media network…
Really, the entire internet now is potentially going to get a little bit of AI assistance and all the millions of people who make the web what it is are going to face a choice: Do I let the machine help me fill this out? And I think that has the potential to change a lot of what we see on the internet.
Well, give us an example of how this could have real world effects. I mean, obviously a lot of people would like some assistance writing. So why do you see this as potentially the end of the human internet?
Well, a lot of the first uses for AI writing have been things where people are kind of stressed out. You know, job applications, writing a tense email to a manager or a landlord or, you know, breaking some bad news and having the AI sort of soften the edges or make it seem more formal or polite. But these are mostly private interactions, and the web is full of people posting in public for other people to see.
So let’s say you’re looking for a review of a product on Amazon. You’re going to read these reviews and you’re going to go through and you’re kind of looking for a little bit of evidence of humanity, like “this is a real person who bought this thing, and they’re telling me what it’s like in their voice, and I kind of believe them.” Now, if you imagine everyone being prompted with this little button, you know, “help me write,” you can sort of see how that authenticity, that human touch and that credibility that comes with being associated with a real person could kind of start to get fuzzy.
The web is already full of stuff that feels like it was written by bots. In some cases it was. And this, you know, isn’t intended to create spam or to be misleading, but it could have the unintended consequence of making people sound a little bit more like machines, or just making people all sound the same. And that really could change the spaces on the web where people kind of make the whole experience worth it – places where people talk to each other.
In your piece, you write “Features like this will be a massive test of how people actually want to use generative AI” and I suppose it’s worth underscoring here that this will be an option coming up. I mean, people could potentially opt out and just write their review on their own, for example, right?
That’s right. And I’ve been using a similar feature that Google’s been testing for quite a few months now. They’ve put a “help me write” feature in Gmail. So whenever I write emails, I’m asked if I want some help. And generally I’ve sort of learned to ignore that feature.
In Google Documents there’s a similar feature where you can start writing and Google will say “would you like some help here?” And you can give it some pretty specific instructions. You can say, “hey, I’m writing an invitation for a party. I want it to be full of jokes.” And it’ll do a pretty good job.
But you generate it, you look at it, you read it back and you’re like, “wait, this isn’t me.” But also it’s pretty good. It’s good enough.
So, you know, I’m discovering in those situations that many times it doesn’t feel appropriate to use. But there are times, you know, where it might feel appropriate to use.
And of course, everybody’s different. Some people might be so relieved to have their computers kind of just write their content for them. And by putting this in Chrome, it’ll really test out what people want from this kind of stuff, what people are really comfortable letting their browser do and letting software write for them.
And we should note Google Chrome is not an unpopular web browser here. We’re talking about billions of users. More than 3 billion users I think, last numbers I saw.
It’s the most popular browser on the planet. It’s the window through which a majority of people use the web see it.
So you know what Google does with their software here could potentially have a huge effect on how people think about AI. This is going to be the first experience for a lot of people who haven’t used tools like ChatGPT or who haven’t been playing around with this new technology. At some point in the not-too-distant future their browser is going to ask them, “hey, can I fill out that review for you?”
Well, is Google playing some kind of sinister role here? I mean, you I can see the argument that if Google didn’t do it, we’ve already opened the Pandora’s box. Somebody else is going to, I would imagine. Or is there is there such inevitability around AI that we need to sort of resign ourselves to that fact and adjust accordingly? Or what would you say to someone who’s trying to evaluate what this means?
I do think the availability of tools like this is inevitable. I think in one way or another, people are going to interact with these things. It’s a new technology. It’s very capable, and it can be very helpful in a lot of situations.
What isn’t inevitable is that people choose to use it. Like I said, I’ve had access to similar features in my Google Docs and my Gmail, and a lot of times it feels kind of strange. I don’t want to respond to a heartfelt email with an automatic response. But, you know, maybe if I’m filling out an application for an apartment or something like that, maybe I wouldn’t mind an AI reading it over and making it a little more polite.
So, I don’t know. Nothing is inevitable about how people use technology, but I think we can say that it is inevitable that we will all try it.