Combating Deepfakes: A Crisis of Trust

Jenny: Hi friends! Welcome to today’s episode of We Are, Marketing Happy – A Healthcare Marketing podcast. My name is Jenny Bristow and I am your host. I’m also the CEO and founder at Hedy and Hopp, a full service, fully healthcare marketing agency. I’m here today to talk to you about the rising crisis of AI deepfakes in video and social media content online, and the impact it can have on us as health care marketers.

This week, the American Medical Association actually introduced a comprehensive policy framework to prevent physicians from unauthorized use of AI-generated deepfakes. So they put out a press release with seven policy principles for a framework to help protect physicians. The AMA CEO, John White, talked about how, with the rising volume of deepfake social media content, online protections need to be put in place for physicians.

And as a marketer, I completely agree. We have all seen social media clips go viral, and to a more discerning eye, we can easily see that these are AI. But more and more often this content is becoming so sophisticated that it’s really difficult to be able to see with the naked eye if something is fake or not. So the AMA framework is built around seven principles.

I will link to a full article in the show notes for you to be able to read it in depth, but it really focuses on physician identity protection, making sure that there’s explicit informed consent if their likeness is used. Banning deceptive impersonations, making sure that consent is opt-in with consent, being able to be revoked. Mandatory labeling and transparency. A shared responsibility around platforms, hospitals and vendors implementing safeguards, including rapid takedown mechanisms and clear labeling if content is AI generated.

Enforcement and remedies if AI content is found to be used, and minimizing the administrative burden and ensuring that protections are the default with standardized consent processes that don’t create an undue burden. So the seven policy principles are meant to create a framework for how, physicians really should be collaborating with hospital systems and vendors to be able to ensure that their likeness is not used irresponsibly, or with a deepfake when, it is identified, you know, how can you quickly rectify this?

So that’s what I want to talk about today is as a healthcare marketer, what is your role in that? I mean, right now, we’re really seeing a crisis of trust for content online. This really started gaining traction during Covid when we were really dealing with misinformation, and now we are really dealing with the ability for technology to impersonate a physician and use their likeness, video or voice only, and making it seem as though they put content out.

And so we’re really having a crisis of trust in this moment. And so, as a marketer, our new mandate is definitely that we need to fight fake video and audio of doctors within our systems. And the antidote to that really, I believe, is undeniable authenticity. Right? So really making sure that we are putting out content that is easily authenticated by viewers and making sure that we’re creating a premium human in the loop experience for our viewers and for folks that are trying to research things online.

So search engines like Google, Google has the EEAT experience, expertise, authoritative ness, and trustworthiness approach to content and content that they prioritize within their search rankings. That’s really the way that we should be viewing the content that we’re putting out for viewers online. Our patient population also really our greatest asset as health care marketers, as our real living workforce.

So how are we going to actionize that? I have a handful of tips of ways that you could be thinking about action, to be able to move some real content forward. So first is you have to realize that raw, unpolished content is the new high-production video. Nobody wants super highly produced content anymore. If you think about social media influencers and kind of the TikTok style of content, it is raw.

Not highly edited, not highly polished. They use, you know, mobile, inexpensive microphones, usually shot on an iPhone. AI videos, deepfakes especially look usually highly polished. They usually use generic stock backgrounds. And so you can combat that with raw, behind-the-scenes, unedited video content. So, for example, physicians could film some or some, in nurses, somebody within your system, you can have them some quick day-in-the-life shorts or record them walking through hallways of your actual recognizable hospital.

Right. Like make the scenes in the background, something that visually, patients will be able to recognize as being part of your system. Also think about verification social campaigns. So actually create a campaign around something like real doctors, real answers. Right? So explicitly talk about how you’re creating verified advice and how you are putting real physicians or clinicians at the front to be able to talk with the patient population.

Use features like Instagram Live, LinkedIn audio spaces, interactive Q&As, anything where you can really go back and forth with folks. And it’s not something that an AI deepfake really could create. Make sure, next, that you have a multi-channel footprint. So don’t just rely on social media videos, which are the easiest to fake, but build an entire ecosystem. Make sure on your physician profiles that you have video content there that a viewer, if they see a physician on another channel, they can go to the physician profile and see if that content matches, see if, you know, the likeness matches.

And then finally localize the content, how the content that you create not be something that can apply to everyone across the country. Talk about hyper local issues like local allergy spikes, community wellness, event specific regional health trends, things that are happening in your back door that a deepfake, you know, just wouldn’t even have interest in participating in.

It will really anchor their authenticity and the physical community that you live and serve. And then the last thing you need to think about is how, as a marketer, can you help your organization comply with that seven step framework that AMA is working to introduce. First, make sure that you are auditing consent and media releases. One of the things that AMA highlights is that physician likeness is a protective right, and calls for explicit opt-in and revocable consent.

Remind your team to review their internal marketing media releases. Are your old forms giving blanket permission to use their voice and image forever? If so, work with HR to update forms on file. Make sure that you’re complying with this new framework, and make sure that you have all of your T’s crossed and I’s dotted. Next, you have to also consider that shared responsibility framework.

So the AMA stated that they believe hospitals and health systems share the responsibility to police deep fakes. So as a marketer at a system, one of the things that you can do is if your marketing team spots a deep fake or copycat account using money, your doctors names are faces. You need a rapid response plan. So work on creating that rapid response plan ahead of time, and then make sure that your team knows how to execute a fast takedown request on meta YouTube, and TikTok and then finally use it as a positive internal communication opportunity with your physicians.

Write a memo or newsletter explaining that you know, AI defects are a rising concern in the medical community. Here’s what we are doing to be able to protect our physicians and combat the fake content that is being put online, and help make sure that our patient populations can access reputable information from our system. It is a massive opportunity for you to build trust within your own system, for your physicians to know that you are, number one, aware that this is something that they are dealing with.

And number two, that you’ve got their back and are happy to not only handle some of the foundational components, like making sure you have a rapid response plan in place, but also if they have additional questions, that you can be a resource to help make sure that they feel educated. Some AI platforms like Gemini are doing some things to be able to make sure that AI generated photos and content.

Video content can be easily identifiable, like Gemini, for example. It’s called a SynthID to digital watermark. So even if you crop out that little star watermark, that it automatically puts on it, that is. And the SynthID is embedded in the code. So it’s very easy to be able to identify if it’s generated by AI. Hopefully all of the other platforms will be following suit, and this will be something that, in five years from now, will be able to look back on and laugh, because it’s something that we’re no longer dealing with.

But until then, being proactive and making sure that your physicians know that you’ve got their back and you’re on their side will go a long way to building trust not only with them, but also with your broader patient population online. So this episode was helpful. Please give us a like, be sure to follow the podcast so you get notification of our new releases each week.

And then if you found this useful, share it with a colleague. If there’s somebody within your organization that you think would find this information helpful, give it a share. Again, we’ll be linking the AMA release in the show notes. You can read more about it in depth. And thank you for tuning in to this week’s episode of We Are, Marketing Happy. We will see you next week! Cheers!