Adobe has announced a plan to
authenticate photos as way way of helping consumers determine if the images they see on the Internet are real or fake. How this will work hasn’t been fully spelled out, but is expected to be rolled out in “coming months” according to Chief Product Officer Scott Belsky (See the video of his presentation
here.)
Allen Murabayashi, Chairman and co-founder of PhotoShelter,
cautions photographers not to get too excited about how this might work. He explains in the article I’ve copied from the PhotoShelter website and printed a little more about what to expect and the likelihood that it won’t really help creators or consumer.
It seems to be a new metadata standard that will capture creator information and track changes in the original file. The immediate problem is that many social media platforms strip the metadata. So if Google is finding images on existing websites it is hard to tell what percentage of them will have any data that can be tracked.
For convenience I’ve printed the key elements of Murabayshi’ analysis below:
What exactly is it?
Details are murky. But the initiative sounds like an idea composed of two main parts:
1. A metadata standard that captures creator information (akin to IPTC) and tracks changes to the original file. This could be some sort of distillation of the History panel in Photoshop.
2. An image registry, which acts as a centralized (or perhaps a decentralized mechanism using blockchain technology) repository of metadata.
Don’t get too excited
On its face, the announcement and alliance seems like a positive development for photographers to address issues of “orphaned” works. But there are a few reasons to be skeptical:
1. Metadata standards for attribution already exist. The problem is that many social media platforms strip the metadata. Sites that don’t strip the data, rarely display it. Google Images took a step in Oct 2018 to display copyright and creator info.
2. Image registries have never succeeded. In the past, the lack of support and adoption by major players (e.g. Google Images) has spelled doom for image registries. Adobe’s support is a huge boost in the right direction.
3. Adobe hasn’t clearly defined their goal. Is attribution the most important (great for creators)? Is authenticity most important (great for news sources, good for creators)?
4. A much broader coalition is needed. Top-tier news organizations already vet their photo sources and provide attribution, so the NYT’s involvement doesn’t do much for photographers or consumers. Twitter’s support is non-trivial but without similar support from Facebook, Instagram and other social media platforms that have been plagued by misattribution, theft, manipulation, etc, the initiative is likely to fail.
5. Amplification of “fake” news/content is unaddressed. Belsky states, “…over time consumers will expect content to come with attribution.” I think this is fallacious thinking. Misinformation will spread even if you label it as such. We’ve all seen friends post satirical pieces only to have their friends and followers take it seriously. Attribution is an important foundational step, but until media platforms disallow amplification of “inauthentic” content (e.g. removing “Share” links or making content less visible in news feeds), misuse and theft will remain rampant.
It’s early days, so let’s hope
Adobe’s motivation for addressing this problem seems sincere, and they undoubtedly have many smart people thinking about the issue. Hopefully they will be able to persuade a critical mass of companies to join the initiative and develop and deploy the technologies needed to make it a success.