TLDR - too long didn’t read/aka summary
Truple has a brand new AI for detecting nudity and sexual content in screenshots. We believe it’s the most advanced AI detection available in any parental monitoring or accountability app.
- You choose what gets flagged. The sensitivity slider is gone. In its place you pick exactly which content is flagged as high risk, by body area, by male or female, and by nudity or sexy.
- Still 100% on the device. Screenshots are scanned on the device itself, just like they always have been. They are never sent to us to be scanned.
- Faster and easier on the battery.
- Rolling out now. It will reach every platform over the next week or two. The new controls appear for a device once its app has updated. Until then, that device keeps the sensitivity slider.
A smarter AI for screenshot monitoring
Until now, you had one control: a sensitivity slider. Slide it up and more screenshots were flagged. Slide it down and fewer were. It worked, but it treated every family and every person the same.
The new AI understands far more about what’s actually in a screenshot.
It recognizes 21 different categories of content.
For every screenshot, it can tell:
- Which part of the body is showing.
- Whether the person is male or female.
- Whether it's nudity, or someone dressed in a revealing way.
That means you no longer have to guess at a slider setting. You simply tell Truple what matters to you.
Choose exactly what gets flagged
Once the app on a device has updated, that device gets a new Content to flag setting in place of the sensitivity slider.
Click Customize and you’ll see four figures: female and male, front and back. Tap any body area to step it through three levels:
- Off – that area is never flagged.
- Nudity – flagged when the area is uncovered.
- Nudity + Sexy – also flagged when it’s covered by revealing clothing, such as swimwear or underwear.
A few examples of what this makes possible:
- Flag women in swimwear, but not shirtless men.
- Flag nudity only, and leave everything else alone.
- Flag everything, for the strictest protection.
In a hurry? The Quick set buttons switch between Default, Everything, and Nothing in one tap. Most people will be well served by the defaults, and you can change them any time.
How to change what gets flagged
- Sign in at app.truple.io and open Devices.
- Select the device you want to change.
- Under Detection, find Content to flag and click Customize. (Still see a Sensitivity slider here? That device’s app hasn’t updated yet.)
- Tap the body areas until they’re set the way you want, then click Done.
- Click Save.
Each device has its own settings, so a child’s tablet can be stricter than a parent’s phone. As with any device settings change, the account owner and everyone who receives reports are emailed when you save.
Private by design: the AI runs on your device
Truple has always scanned screenshots on the device itself, and that isn’t changing. The new AI runs entirely on the phone, tablet, or computer being monitored. Your screenshots are not uploaded to our servers to be scanned.
This is what allows AI scanning to work together with end-to-end encryption. You get smarter detection without giving up any privacy.
Faster, with less battery use
The new AI isn’t just smarter. It’s also faster and uses less battery than what it replaces, so the monitored device feels the difference too.
When will I get it?
The update is rolling out over the next week or two on every platform we support.
The new controls are only available once the app on a device has been updated to a version that supports them. Until then, that device keeps the Sensitivity slider and works exactly as it does today. Once the app updates, the slider is replaced by Content to flag.
Because this happens device by device, you may see the new setting on one device and the slider on another for a little while. Keep your apps up to date and every device will get there.
It will keep getting better
Accuracy is significantly improved, but it still isn’t 100%. No AI is. At times a harmless screenshot will be flagged, and at times something will be missed. We’ll keep updating the AI so it becomes more and more accurate over time.
You can help. If you notice something being flagged that shouldn’t be, open a support ticket and let us know what type of content is being wrongly flagged.
As always, risk ratings are there to help you focus your attention, not replace it. No algorithm can match a trusted friend or loved one reviewing the screenshots, so please keep reviewing them as often as you can.
Cam