User data is too valuable, it is very tempting for tech companies to try all means to collect user data. The privacy cost of modern tech consumers are becoming higher. As product creators, we wanna challenge the temptation by putting privacy first in developing our new product. To do so, we have technical and business compromises supposedly to make or find better solutions for. We don’t know if it is better off for our users, as well as for us ourselves. Or we just won’t win convenience and economics. We’ve just started, and I’d like to share the journey.
Our team is working on video recognition, and we are excited about its application as a security camera. We set out to bring the smartest features to a security camera. But as we research the market, we found that privacy is a huge concern, especially for AI cameras. It makes sense because:
- Personal footage is highly sensitive data. People are insecure about the idea that their living room activity or their baby’s image is hacked. Video is powerful data, which hackers have more incentive to spy.
- Cloud servers are not entirely safe even with encryption.
The smarter technology becomes, the more we are concerned about privacy. And it is never a simple issue to discuss. In this matter, however, we think we can take the bold step. This is what we decided to do:
- Proceed video offline. Streaming feed is not uploaded to any server for AI processing. (Note: recorded footage could be stored in chosen cloud service upon user’s choice).
- No profile required. We take no interest in user’s identification information, nor link to third-party apps.
- No in-app tracking. No analytics. We do not follow our users.
Compromises we have to consider
The reason businesses find it hard to put privacy first is because it comes at various costs.
No user data = worse product?
AI smart features are the core value of our security camera. AI image/video recognition needs ongoing data update to improve accuracy. Does not collect footage mean we have to compromise our product quality? We don’t think so. AI is increasingly open. We can use public data to train our algorithms to achieve satisfactory quality. Collecting user data for training purpose could just be an excuse sensibly communicated to you via marketing efforts.
No user data = less convenience?
Without usage data, we might not be able to implement automatic personalization features. But let’s think again, is it the only way to create more convenience for the user? We aim for good user experience created by design and giving the user control over how to use the product.
No user data = better user understanding?
We won’t be able to first analyze user demographic information, and second see how users interact with our app. However, we sustain a user community who we actively listen to. We are also self users, so we design and develop a product exciting enough for ourselves.
No user data = less appropriate marketing?
Since we don’t track in-app activities, we won’t be able to analyze the effectiveness of paid marketing campaigns. But we believe quality marketing comes first in thoughtful communication, which gets attention. It helps us worry less about following our users.
No user data = loosen user relationship (lower retention & less lucrative biz model)?
It sure would be harder for us to maintain interaction with users when we don’t use in-app messages. It would also seem more difficult to ‘lock in’ users with no binding cloud subscriptions. After all, however, we think what matter is how often users actually use our product, which is an effect of its value to users.
Actually, our modern life is already overly connected, isn’t it? We think a bit less connection does us some peace of mind and more quality living.
Wanna join a hand?
We are still figuring things out. If you think we should do what we are doing, you can support by:
- If you are in our market, be our beta users and give us feedback.
- Spread the words to more privacy-concerning folks or those you think should care about the topic.