Ben Semmes is the chief operating officer at Monotype.
Generative AI is omnipresent in the news and on our social media feeds. People are both excited about its promise and worried about its potential misuse.
Today, users can utilize generative AI to manipulate existing media or even create new content with just a few strokes of the keyboard, generating essays and email campaigns; modifying images, audio or text; or even generating 60-second photorealistic videos based only on text instructions. From Taylor Swift deepfakes on X to AI-generated headshots and dating profile photos to manipulated political disinformation on social media, generative AI holds great disruptive potential.
How can content consumers trust that the videos, photos, audio and text they see are authentic and haven’t been manipulated? How do we, as businesses, maintain trust with our customers and collaborators? How do we empower our customers by providing trust indicators that communicate and reinforce the authenticity of our content throughout our design process? How do we harness the potential of AI while side-stepping its potential hurdles?
Making An Authentic Brand Identity
Customers’ trust in your business is key to your success. Having a consistent, reliable brand identity is essential, helping your customers identify your brand in a potentially crowded marketplace. A strong brand identity also ensures your customers are less likely to be tricked by imposters, keeping their data (and trust in your brand) secure. From your brand voice to your font choice to your brand colors, all of these elements help your customers identify you.
How To Prepare For What’s Next
How do you protect your company’s assets in the generative AI era?
Part of the problem, according to Fortune’s Sage Lazzaro, is that we’re “building the generative AI plane while flying and dealing with the fallout in real-time.”
There are a few main concerns you should prioritize while preparing your creative assets for the generative AI era. First, how can you ensure your assets are properly and consistently attributed to your company? Second, how can you engender trust by communicating the authenticity of your assets and the provenance of any changes to them in a way that is accessible and easily understood? Finally, how can you clearly communicate that your assets are not to be used by AI training models for machine learning and protect your intellectual property?
While we may not be able to perfectly prevent all AI-powered fraud, that doesn’t mean there aren’t steps we can take now, and plans we can make for the future, to decrease our risk.
Proving You’re You: Helping Your Customers Identify Your Content
Generative AI is making it easier and easier to believably impersonate individuals and businesses, with some businesses losing tens of millions of dollars to deep fakes. How can you protect your business and your customers?
One action you can take is to proactively authenticate your profiles and content. First, your brand needs to be consistently identifiable across mediums. This means your website URL, social media handles and email addresses should all be similar and clearly identifiable as part of your brand and company.
Try to claim any handles or URLs that might be easily confused with your brand, so that potential imposters can’t set up shop there. Your accounts should also refer to each other across platforms so that your customers can clearly trace your accounts back to you. Be sure your accounts are also verified.
Creating Content Your Customers Can Trust
You want to make it as easy as possible for consumers to quickly authenticate your content as legitimate.
This can be accomplished smoothly and consistently by applying something called a “trust indicator” to all of your content. Trust indicators are easily identifiable marks and metadata labels that can be used by anyone to quickly establish trust in a digital asset. This will help your customers differentiate between your content and an imposter’s. One way you can label your content is through Content Credentials based on the Coalition for Content Provenance and Authenticity, or C2PA open technical standard.
Content Credentials provide a verifiable way for creators of images, audio, video or fonts to label their work as theirs, as well as any additional information they may want to disclose on the file’s history, including the who and how of any changes made to the asset along the way. Major companies, such as Adobe, Microsoft, Google and OpenAI, are leading members of the coalition and are committed to adopting the open standard, meaning images created in Adobe Firefly or by ChatGPT and DALL-E 3 include Content Credentials based on the C2PA standard.
At Monotype, we have embraced the C2PA technical standard and are working toward staged releases of fonts that embed Content Credentials. In our industry, we believe this technology will simplify the font identification process, helping anyone distinguish between generative AI fonts and carefully crafted human fonts while also protecting attribution for font designers.
We believe this technology has major implications across industries. Content Credentials are linked with the file forever, digitally signed. This information is easily viewable by anyone, not just content experts, making it much easier for customers to feel assured the content they are purchasing or interacting with is legitimate. By giving customers information on the history of a file, Content Credentials let consumers make more informed decisions.
Protecting Your Assets From Scraping
Another way you can minimize the risk to your brand identity from generative AI is by labeling your assets to indicate they should not be used to train AI. Content Credentials allow you to label any content file type with a “do not train” label. There are also some tools you can use that make it technologically more difficult for AI scrapers, such as PhotoGuard or Glaze, which work by making small changes to the images so that if AI scraping does occur, the image data is distorted.
It may seem overwhelming to keep up with the challenges posed by generative AI, but by taking a few of the steps above, you can help protect your company’s assets.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
