Worried about the vast spread of misinformation of AI Images?
To help prevent the massive spread of misinformation, Google has announced an
invisible permanent watermark on the images generated by Artificial Intelligence.
The technology, SynthID lets you embed the watermark directly on the images
created by the platform “Imagen”. Imagen is the latest tool that generates AI
images from text. The best part about this news is that even after modifications like
altered colors and added filters, the watermark remains intact. This tool SynthID can
not only embed the watermark but also scan the incoming images and identify the
likelihood of them made by Imagen. This is done by scanning the three levels of
certainty: detected, not detected, and possibly detected.
Google on Tuesday Aug 29, 2023 wrote in a blog post, “While this technology isn’t
perfect, our internal testing shows that it’s accurate against many common image
manipulations.”
Altered photographs and misinformation:
Tech companies are worried about finding a way to figure out manipulated images
as in recent months, an AI generated image of USA’s former president Donald
Trump getting arrested went viral and Pope Francis in a puffer jacket was widely
shared. Google has thus announced joining growing number of startups and big
tech companies that are trying to find solutions of such issues. Big Companies
including “Truepic” and “RealityDefender”, want to protect the essence of
difference between real and fake. Google has shown its massive interest in this
quest, similarly, The Coalition for Content Provenance and Authenticity (C2PA)
which is an Adobe-backed consortium has been a leader in digital watermark
efforts.
With the AI technology developing faster than humans trying to keep up, it’s also
clear that these technical solutions will be able to solve these kinds of pressing
problems.
