4 November 2020
From basic photoshopping to AI-generated deepfakes, it's becoming ever harder to believe what we see. One solution is battling AI with AI to spot and flag fakes once they spread online. Another is to start at the source, adding invisible watermarks or other proof of provenance when a photo is taken. Chipmaker Qualcomm has partnered with verification platform Truepic to do the latter.
A prototype mobile device integrates hardware (Qualcomm's Snapdragon chip) and software (Truepic's Foresight system) to create photos with cryptographically-sealed provenance data whose authenticity can be verified by recipients. When taking a picture, a phone user simply selects secure mode. Data about where and when the photo was taken is automatically added, along with a way to track if pixels have been changed. Website and app builders can then implement features to read that data. AutoTrader could let buyers verify that secondhand car photos are real and undoctored, and Bumble could do the same for profile pics.
As we highlighted in our September issue of Make→Shift, a fight for facts is on: the more divided people become, the more likely they are to believe their own version of the truth. And as fake and doctored photos and videos become more common and more convincing, it's not much of a stretch to say they pose a threat to democracy. And to commerce ;-) Which presents a wide range of opportunities for entrepreneurs willing to help consumers tackle digital misinformation and rebuild trust.
The TrendWatching content team