Learning from our mistakes

Originally published at: http://mycroft.ai/blog/learning-from-our-mistakes/

On Thursday May 3rd our marketing team sent out a survey intended to help us better understand Mycroft’s users. Shortly after it went out, several vigilant members of our community pointed out that the tools used for the survey included third party tracking software from companies with less than stellar reputations for privacy.


The survey also included questions about ethnicity, income and national origin that some users understandably found offensive. The questions were worded in a way that is acceptable in some parts of the United States, but are clearly unacceptable in a global context. Our team was thinking locally when we should have been thinking globally.


Those of you who’ve seen me speak in person or on video may have heard me say: “We are not going to be perfect. Creating systems that are truly private in a world full of surveillance is difficult, but I will tell you this - systems or software that compromise your privacy will last only so long as I am unaware of them.”


I had a sit down with the team involved and helped them to better understand what we did wrong and why it is out of line with our ethos. This is something I should have done proactively and I apologize for not doing so. I let them know that going forward they need to keep one concept at the forefront of their minds when interacting with users and user data:

"It is always the user's choice whether or not to share their data with a third party."


The responsibility for our failure rests entirely with me. I am accountable. I am accountable to our users and I am accountable to our employees. I am accountable for making sure they understand our ethos and live up to our core values.


I’ve instructed the team to irrevocably destroy all of the data collected during this effort and initiate a comprehensive review of our online presence to identify tools or systems that may compromise user privacy. Though we do need to understand our users, we need to do so using tools that let them choose whether or not to share their information.


I let you down. I apologize. I’ll work to do better.


Joshua Montgomery


Mycroft AI


Thank you for this statement.
I also received this mail and startet the survey as i wanted to help you understanding “us” as users better.
But i left the survey after the 3rd or 4th question as i found these kind of questions unacceptable.
As we are all humans we all make mistakes or get a misunderstanding position.
But your apologize for me is a big gesture and shows me that MyCroft staff is on a good way.

Best regards


It’s good to see this Human side of Ai, and this is why I support the Mycroft ‘journey’ - it’s a project that deserves to succeed because it understands the human element of Ai, be it success or failure, and not success at any cost (read data slurping, or underhand data selling) - for this immature technology the trust built in the community is paramount, and when things go wrong (as they always will), it’s as much, if not more about how the problem is dealt with as the problem itself.

So, thank you for the open, honest statement, I support Mycroft in all it’s future endeavours.

1 Like

Excellent response! Honesty and accountability - can be rare attributes in our modern world. Yes, this is such a global project it’s good to remember at all times that the community spans the globe!

1 Like

Thanks for this honest statement. Glad to see this, this is way beyond what i expected to see and its good. This builds trust. As mentioned - nobody is perfect, but we can always try to get better.
Thanks to the whole MyCroft team and to Joshua :slight_smile:

Greetings from the facebook guy from europe!

1 Like

Nice to see how to properly handle a misstep, so rare these days.

1 Like