DAM high standards
keeping up with new technology while maintaining metadata excellence
No matter how much time passes, some of us (ahem, myself included), never quite get over gifted-kid syndrome. If you too were shuffled off to an advanced Math or English class, fear trying things where you aren’t sure you’ll excel quickly or hold yourself to higher standards than you would others, welcome! I would argue that the pros - from academic success, a thirst for knowledge, and great curiosity - make up for the cons - anxious procrastination, difficulty letting go of a project, and missing out on trying something new - I’m working on decreasing the negative aspects as I grow. One area where I don’t think I’ll be conceding much, though, is the world of standards.
Standards are the rules that keep our industry (and I’d argue only slightly facetiously, the world) working. Knowing the rules can help with decision-making and even knowing when the project is successful. From naming conventions, asset structures, and tagging strategy to work processes, standards can answer questions once and allow us to move on to the next problem to solve. Striving for excellence is often enough to create wonderful work together.
If there is an organization that takes standards even more seriously than I do, it’s the International Press Telecommunications Council (IPTC). In response to how GenAI continues to grow swiftly, they have shown how our systems can adapt and grow too. The IPTC NewsCodes Working Group has recently added a new digital source type to cover ‘inpainting’ and ‘outpainting’ in generative AI tools like DALL-E and Adobe Firefly. Now we can call this type of asset a Composite with Trained Algorithmic Media and credit the original artist, the digital artist who prompts the tool, and the GenAI tool that generated the output.
Why does this matter? GenAI is barreling ahead in every industry with a variety of outputs from writing, coding, customer service, editing, illustrating, and more. But if we want this technology to improve how we work, we are going to need standards.
What fun is a board game with no rules? A race without a finish line? Some people might say the journey is the point, but I also enjoy the reward at the end of the journey. Other than a sense of satisfaction, without rules how can we be fair?
Fairness is probably my top core value (take 30 minutes and figure out your own, it’s magical and I just saved you $1500 on a career coach) for this reason. I am a big believer in everyone having to follow the same rules and nothing angers me more than seeing cheaters win. What is fair can vary based on the situation, of course, but not that much. When there is an agreed-upon set of standards that applies to all participants, it can be much more clear when the situation is actually fair.
Back to GenAI content, where the standards are still being written. How can they improve the situations where we currently see conflict - from privacy, intellectual property, bigotry, and access - without stifling the benefits? I think each industry is going to have to respond quickly with its own standards. For example, in the writer’s and actor’s strikes now happening, their unions have prioritized the impact of AI data collection and replacement of jobs. The idea of an actor being paid for a day of work where their likeness is captured and reused forever is bleak. But maybe a GenAI tool can come up with 10 possible dog names that make sense for the character to choose, then the writer can select their favorite. Getting rid of the drudge work is what other technological advances have offered. Maybe there is a way to minimize the negative fallout.
In DAM, we have benefitted from AI that helps crop images and video by recognizing focal points and adjusting to fit the needed dimensions and other AI tools for suggesting keywords to populate metadata fields. As we expand what an asset even is, I will be looking for the tools that remove repetitive tasks and allow creative, marketing, IT, and content teams to adjust their workflows to focus more on what humans do better than robots. A common language with a shared process has been benchmarks of knowledge-based systems from the Dewey decimal system to zip codes and there is no reason to stop now. What the IPTC has been doing will need to be repeated across our industry to keep a common set of standards, well, standard.