Not even Google can take your data on Nextcloud

On July 1st, Google updated their privacy policy and it’s struck a deep chord on the Internet.

Claims like:

“Google is using your Google Docs to train AI”

“Bard is trained on your scraped web data”

And even, “All of the internet now belongs to Google’s AI” have flooded data privacy and technology news outlets, shaking the concerns of creatives, professionals and frankly any individual who cares about their personal data.

Nobody wants their work, ideas or words taken for granted and used for a company’s or someone else’s advantage.

So in light of it all, what’s the truth?

A shift in Google’s privacy statements

In its privacy policy, Google added a clause that any ‘publicly available data’ will be used to train its AI tools like Bard. According to several sources, the statement used to say that data will be used “for language models” whereas now it’s been updated to “AI models.” The mention of tools Bard and Cloud AI have also been added in addition to Google Translate.

As Jack Wallen points out at ZDNet, it’s rather unclear what that means exactly for your private documents on Google Docs or Workspace. For instance, when you don’t share a document at all, is that private? What if you share a document with one other person? Or make it a public link that you share by email to your friends or colleagues? Is that public? Without a doubt there’s some definitive grey area, and frankly, is it worth the risk? Not to mention that the privacy policy can be changed again at any time.

What does this training on your data mean?

There are a some risks and unethical issues. First of all, copyright issues, with AI simply repeating what it learned. For instance, voice actors‘ jobs and raw talent are at risk due to AI copying the „likeness“ of their work to produce more of it unauthenticated. In another case, GitHub Copilot is plagiarizing the code of copyrighted coders work. AI is on the verge to replace people’s raw talent with an abundance of unauthentic work that the masses can utilize. Second, your data is at the hands of companies and groups who’s only goal is to have the best AI model out there. With your data up for grabs, it means it’s out of your control even if it’s genuinely yours.

On Nextcloud your data is never leaked, used or mishandled

At Nextcloud, we can guarantee you with 100% certainty that your data would never be leaked or misused to benefit a third party – especially to train AI models.

The reason we can be so matter-of-fact is because of our self-hosting options. At Nextcloud, you as a user can choose to host all your data on your own server(s), which allows you full reign and control over your data.

At times like these when platforms like Google provide discreet policies around data privacy, you need a reliable service you can trust.

At Nextcloud, self-host your Office space, Mail and Contacts work environment and files or use a trusted hosting service that we partner with to do it all with less technical know-how and effort. Either way, your documents, files and other valuable data won’t be at the disposal of prying big tech companies.

Nextcloud AI: An ethical, on premise approach

If you’ve been following Nextcloud News, you’ve probably already heard about and even installed our new AI features. The newest additions to our Hub 4 and now Hub 5, these AI tools are designed to be locally installed, meaning the data you’re interacting with using the model doesn’t leave your cloud.

AI can and should be implemented and used responsibly, which is why Nextcloud offers ethical and on premise AI tools that give users the opportunity to work with them. So far, Nextcloud provides text generation, image generation and speech-to-text dictation that all can be self-hosted and installed via our Local AI app. Also available are features that are exclusively trained on your server, for example the Priority Inbox. This smart inbox intellingetly puts your most important emails on top of your inbox. Same with our suspicious login detection that uses your past login data as a tool to detect an intruder so you can act in time.

For our new translation AI tool Nextcloud Translate which works in Talk and Text documents, we use our own data to train it and never thought, let alone could, use yours.

The true benefit of using Nextcloud’s AI tools is that you can use your data to work towards training an AI model, if you choose to. That entails that everything you do on your Nextcloud platform can be used to improve and make more efficient your tool, and yours only! That’s the difference and the way it was meant to be.

For all the details on our AI features, integrations and apps, check out our latest release blog and stay tuned for more blogs about it!

Note that all integrated AI features can be disabled and all AI integrations and apps are entirely optional to install and use. If AI is not your thing or you’re not interested at the time being, that’s totally okay and accepted. Hey, it’s your cloud, we’re just here trying to give you as many cool new options as possible!

Don’t take the risk

Whether Google is misusing your data from Docs, Sheets or other public information like on your blog or social profile or not, we’d rather not take the risk.

Join the new era of AI, with Nextcloud!

We can’t wait to see more users using AI in a secure, controlled way.