rhindi

Hey! We are publishing the whitepaper in a few weeks. We use a blockchain in a number of places, both on the consumer side for the skills marketplace, and on the developer side for federated learning.

Speech to text is precisely the most sensitive thing to protect. There is simply no way to offer real Privacy when processing voice in the cloud.

The french privacy council just issued an official warning regarding cloud-based voice assistant. They recommend switching them off when not using them, and telling your friends you have one when the come over!

Not sure they will be GDPR compliant if it continues like that..

https://www.cnil.fr/fr/enceintes-intelligentes-des-assistant...

We will be open sourcing it gradually over the year

Every component publishes a message on the local network, so you can have other devices like an iPad capture the output

Actually, we have an example of this happening with messenging apps. No one used to care about end to end encryption, until Telegram and Signal became popular and WhatsApp followed up. The result is that today, consumers demand it in any messenging app.

Hey author here!

I think the debate comes more and more often in the public opinion, and I believe privacy will soon become something people will demand when they buy hardware and software, just like memory capacity or battery usage.

But we don't need to wait for politicians to move on this issue, we can already do a lot without tem! There is growing research on privacy-by-design algorithms, on which we work at my company Snips (https://snips.ai), which will ultimately prevent governments and companies to use data without having the explicit permission of users, what we believe is their right!

Author here!

I am the founder of https://snips.ai. At Snips, we are building an assistant that is private by design: what I describe in the article are things we actually work on today :-)

Happy to share my thoughts on it and contribute to the debate!

More details about what we do:

Like the big software companies (the G, the A, the F, the M), we want to build an AI to help you get what you want more effectively.

But we believe this can only be done if we take care of people privacy: your AI should ultimately know your favorite restaurant, your girlfriend’s name, but also your health record and everything else you might not always feel comfortable sharing with the world at large.

This is why an AI alter-ego will have to work with privacy-by-design. You should not have to trust any company whose CEO might soon be replaced by another not as trustful one.

We think about your data as your most intimate and valuable property, and we want to use it without compromising its safety!