Gmail’s New AI Era Begins  ## Metadata - Author: [[Hard Fork AI]] - Full Title: Gmail’s New AI Era Begins - Category: #podcasts - URL: https://share.snipd.com/episode/e87dfcff-e273-41d1-a90b-65543e7c971a ## Document Notes ## Highlights - **Host Is Bullish But Critical** - Jaeden expresses strong enthusiasm for Google's AI advancements, especially Gemini and transformers legacy. - He pairs this admiration with willingness to call out practices he finds unethical. Transcript: Jaeden Schafer Everyone knows that listens to this podcast. I am very bullish on Google right now. They're pulling out a ton of incredible advancements in AI when it comes to Gemini. Of course, they're the ones that originally founded the entire AI industry with the transformer paper and the research that they helped complete there. So Google is not going anywhere anytime soon. And typically, I'm quite positive on everything they're doing. ([Time 0:00:00](https://share.snipd.com/snip/5fcd8f19-8c9b-493c-a8e4-6db0692d83f9)) - **Gmail Feeds AI Training For Gemini** - Google is using Gmail content and attachments to improve Gemini and other AI features behind the scenes. - This data access is framed as necessary to power personalized inbox features and smarter replies. Transcript: Jaeden Schafer So Google has kind of stealthily added a bunch of features that allow Gmail to access all private messages and the attachments that you have in your Gmail to train their AI models. I think this is an important story to listen to because this does also impact a lot of other companies that may be looking at doing this as well in the future. So the reason behind this is essentially the fact that Google's AI assistants like Smart Compose, which essentially help you do these AI generated replies in your Gmail, you've probably Seen them. They have Gemini built in now. So when you're reading your email, there'll be a little bubble that pops up, which is this AI-generated reply. It's super useful. And I think a lot of people are like, okay, yeah, Google obviously has to read my emails in order to understand how to do a reply. That's not a problem. I think what people are concerned about is the fact that Google could then take the contents of your email and train Gemini with them. And Google's like, well, we want to make Gemini better, more useful for everyone. But I think that is a line that a lot of people don't feel comfortable with. The reason behind all of this is that Gmail needs a lot of new features to power its Gemini AI. We've seen just Gemini 3.0 come out recently. It's got a lot of incredible features in there. They're trying to help it become more faster. They want it to be able to manage your entire inbox more efficiently. And to do that, they need real email content. ([Time 0:01:43](https://share.snipd.com/snip/18d5abe1-6391-4f0a-ba0d-70dbbb18f38c)) - **Privacy Promises Don't Erase Trust Issues** - Many users feel uncomfortable with private email content being used to train models even with anonymization and security promises. - Google claims strong privacy protections, but default opt-in undermines trust for some users. Transcript: Jaeden Schafer But I think that is a line that a lot of people don't feel comfortable with. The reason behind all of this is that Gmail needs a lot of new features to power its Gemini AI. We've seen just Gemini 3.0 come out recently. It's got a lot of incredible features in there. They're trying to help it become more faster. They want it to be able to manage your entire inbox more efficiently. And to do that, they need real email content. And they also need attachments, right? They need to understand not just what's inside of an email, but, you know, like how, what attachments and other things are kind of attached to it as they're working on getting Gemini To completely manage your email system. And to not just train but also refine their AI models. So a lot of users are now reporting that because of this there are settings that are switched on by default which are instead of just explicitly asking you to opt in and this is where I would Say I'll probably call it out sure if Gmail is like hey can you switch on the settings so we can like read and all your stuff and train on it if that's cool with you and then we'll you know give You access to the best Gemini model. If that's what people want to do, fine, that's great. I do not like it when these models by default switch stuff on. I've seen this happen a couple other times. Google is a culprit frequently, but there's others as well. I think if you don't manually turn these off, your private messages might be used for AI training behind the scenes. And even though Google promises, like Google's not just like, hey, we're going to take your emails and now everyone in the world is going to read them. Like that's not what they're saying in their terms. They promise really strong privacy measures. They have anonymization. They have data security during AI training for anyone handling sensitive or confidential information. ([Time 0:02:46](https://share.snipd.com/snip/c0573709-6bfd-4860-8a6c-609b5e841424)) - **Weigh Features Against Security** - Decide based on priorities: keep the smart features if you value AI functionality, or turn them off if security matters more. - Be aware that opting out removes access to certain features like Gemini-powered suggestions. Transcript: Jaeden Schafer So, I mean, maybe that's the shocking take. Now, if I worked in, you know, security, and I didn't care so much about the AI advancements, but I really care about security, I'd probably turn it off. I don't know. So everyone will fall into different into different camps here. ([Time 0:05:22](https://share.snipd.com/snip/ee35799b-ae57-4632-babc-58ed988bd256)) - **How To Opt Out Of Gmail AI Training** - Do disable two separate settings to opt out: Smart Features in Gmail/Chat/Meet and Google Workspace smart features. - After toggling both off, hit save and refresh to confirm they remain disabled. Transcript: Jaeden Schafer But here's how you turn it off. It essentially acquires requires you to change settings in two different places. The first thing that you do is you turn off smart features in Gmail chat and meeting settings. You basically go to you have to do on the desktop or mobile app, you click on the gear icon, see all settings, you go over to settings and you find there's a section called smart features. This is in Gmail chat and meet, you got to scroll down quite a bit to actually go find that but inside of that smart features you uncheck the option so essentially in what it's going to say It's like if you want to turn it back on what google says is turn on smart features in gmail chat and meet when you turn on the setting you agree to let gmail chat and meet use your content And activity in these products to provide smart features and personalize your experience okay and there's like a little learn more button where you can learn more about what it everything That's actually entailed and beyond just like we're using it to personalize. Okay, but they also like literally using it and training models with it. So there's that. Another thing you need to do is turn off Google Workspace smart features. So still inside of settings, you go to Google Workspace smart features, you click on manage workspace smart features and settings, you're gonna see two options, smart features in Google Workspace, and also smart features in other Google products. Both of those have to be toggled off, then you gotta hit save for that actually to go into effect. Then you want to check if both of them are off, make sure all of the toggles remain off if you refresh it. ([Time 0:05:36](https://share.snipd.com/snip/8fb45421-3f69-4dba-b76e-f2bad6fa9689)) - **Fragmented Settings Lead To Hidden Opt-In** - Google split smart-feature controls across multiple settings, making full opt-out non-obvious and cumbersome. - This multi-location approach increases the chance users remain opted in by default. Transcript: Jaeden Schafer Your email chat and meet with smart features used across all of the other apps. So to fully opt out of feeding your data into AI training, both of those have to be disabled. I will say your your account might not show those settings enabled by default yet. I was looking at a report by someone who said that theirs did not. But if you care about security, I would double check on that. So I've said it before, and I'll say it again, I love all of the new AI features. And personally, I'm actually not going to disable this because I care more about the features and functionality than I do about the security aspect of things. So I'm going to keep using these settings, but I know there's a lot of people who feel rubbed the wrong way by the fact that this was default opted in for a lot of people, and they don't just Don't want the smart features to begin with. So if you fall into that camp, this is what you can do about it. Overall, I think I'm really excited about everything Google is doing. Not excited about getting opted in by default to having your data taken. But this is kind of the nature of a lot of these AI companies. And this isn't just Google. There are so many companies that are having to do sneaky things where essentially they have access to massive data libraries. Gmail is probably one of the biggest. And they have a very tricky time with security being able to find a way to monetize and actually use that data for AI. So they come up with these sneaky little opt-in things. They change their terms of service and their privacy settings. And they probably made me at some point check a box that said, I allow Google to change how they handle my data whenever they decide. ([Time 0:07:09](https://share.snipd.com/snip/1626d725-cf17-4383-b32a-d4549eacb2e7)) - **Gmail Is Part Of A Bigger Data Trend** - This Gmail situation reflects a broader industry trend of companies monetizing massive personal data pools for AI. - Jaeden warns similar 'sneaky' opt-ins and terms-of-service changes will recur as AI becomes more valuable. Transcript: Jaeden Schafer And this isn't just Google. There are so many companies that are having to do sneaky things where essentially they have access to massive data libraries. Gmail is probably one of the biggest. And they have a very tricky time with security being able to find a way to monetize and actually use that data for AI. So they come up with these sneaky little opt-in things. They change their terms of service and their privacy settings. And they probably made me at some point check a box that said, I allow Google to change how they handle my data whenever they decide. And so, like, technically, we're all opted into this stuff or we agreed to their terms of service. But still, that being said, I think you need to be transparent with your users when you're going to take their data. This doesn't just apply for Google. This is a lot of different companies that are sitting on massive piles of data. So this is an interesting problem that this is not the last time you're going to hear about. We will continue struggling with this problem forever, essentially, now that AI is worth so much money. ([Time 0:08:10](https://share.snipd.com/snip/12e5eee2-bcb8-4287-8aef-7458a7151a1a))