The Misinformation Gap: Why AI Literacy Has to Be Part of Digital Inclusion Work

AI Readiness Starts with Knowing What's True

Very few people were prepared for the explosion of AI services and tools. More importantly, these technologies have been created in secret in an attempt to gain competitive advantages. As a result, few industries — and even fewer social impact organizations like ours — had a chance to help shape this technology before it rolled out everywhere. So now, it feels like everyone is trying to catch up to a runaway train — to get ready to integrate AI into their work and lives.

It seems to us like there are two tiers to AI readiness. 

The first tier is gaining internet access and the equipment to get online. But does someone actually know how to vet what you're seeing online and/or how to use AI tools like ChatGPT? That's the next level up.

You can't possibly be ready for AI if you don't have the basics of internet access, the equipment to get online, and the skills to utilize it. 

Foundationally, digital inclusion has to come first, paired with digital skills that include AI — or people are automatically left out of AI’s potential benefits.

The second tier of AI readiness includes understanding that everything AI spits out might not be true. AI generates misinformation and disinformation — sometimes accidentally, sometimes by design. People encountering it for the first time may not know that. You have to understand how it works to know that what you're getting needs to be taken with a grain of salt.

That's a big ask for people who may not yet be confident about anything related to computer use and security. When they're asking AI questions and it's giving them answers that can seem very authoritative — and might not be — the risk is more than confusion. It's that people act on potentially false information.

Teaching people how to get online isn't enough. Teaching them how to recognize and avoid disinformation and misinformation has to be part of the work of digital equity organizations.

The Divide is Widening

Let’s say an organization jumps to AI for customer service. If somebody doesn't know they're not engaging with a person, or they don't understand the proper way to write prompts, or they don't grasp elements of how AI works — there are potential problems. They have to understand some of the complexities that come with AI in order to utilize it safely and access it as a tool to achieve what they want to do.

Think about it. Lots of people are getting laid off, and some companies are pointing to AI as one of the reasons. And people in the AI space keep saying that AI is not going to take everybody's job. You just need to learn how to use it to be more efficient in your own job — that learning to use AI is the best way to ensure job security. But if you don't already know how to use it, you’re falling behind. You probably can't find a job. And if you do find a job, you can't utilize AI to help you keep it.

People who have AI now and are using it now — and have been using it for a while — are turbocharged ahead of all the people who aren't even online yet.

AI is an accelerator that some people have and others don't.

AI is Arriving Without Warning or Consent

People don't have to go looking for AI. You go into an internet browser and there's already a prompt asking if you want AI to help you. For someone getting online for the first time, that's more than an invitation. It's a source of confusion that lands at the very beginning of their experience — before they've learned the basics.

The use or mis-use of data is equally concerning. AI is now being used in some medical offices for note recording, with no real ability to opt out. The choice is either let them do it or switch doctors. In some doctor’s offices, you can't always reach a real person if you aren't an established patient —an annoyance when one has questions or is trying to make important decisions about their own care. That’s impractical. 

Saying “It's HIPAA safe.” is not an answer to the question about what’s happening to your data with AI. It's probably going into an AI agent. It might be going into the AI’s data pool for learning.

And anyone who has ever read a technology services agreement knows that it references many other companies. That's your health information. When you think about how many times people's identities have been stolen — and you add to that all the information going into AI and how it can be compounded — it is a real concern.

When AI is Used to Deceive

Some of the digital equity organizations we work with serve seniors. They tell us stories of identity theft and people losing money because they've been targeted by an AI-based scam. Either voices or the script is right, or information is pulled and utilized against these people and they fall for it. Or they don't know who they're sharing information with. They think they're sharing information with a person they're engaging with online, and it's actually some kind of AI bot doing nefarious things with their information.

Disinformation risk is built into that moment. 

AI-generated content — text, voices, personas — can appear completely credible to someone who has no framework for questioning it. For new internet users, there is often no instinct to ask: is this real? Is this true? Did a person write this or say this?

This is what the fraud looks like now. It's no longer just phishing emails. There are avatars that can call you and sound like real people — sound like your grandson, based on posts he put on the internet — and interact in real time. It's protecting against a new kind of fraud that is exacerbating all of the existing fraud methodologies, creating new ones, and making them more effective for the fraudsters.

The numbers we’ve heard are cause for concern. In one region whose work we follow closely, elder scam losses totaled about $1.8 million for all of 2025. Five months into 2026, they've already matched that number.

A lot of the organizations and other groups we talk with are finding that safety is what's driving AI readiness initially — in terms of how they're starting to incorporate it into their trainings. It’s clearly top of mind.

Digital Navigators Are the Linchpin — And it Isn't Scalable

Digital navigators are increasingly taking on the role of teaching AI literacy. They're teaching classes and educating as they go when they're working one-on-one. They tend to be where people start and building on top of a thin foundation.

But digital navigator work is inherently one-on-one, and that means it doesn't scale easily. A lot of times they end up having to customize. Someone might need to know how to set up an email address before they can even think about setting up an account through an AI platform. They need to have some of those baseline pieces first before making the jump into AI usage.

Could AI itself support the work of digital navigators?

Digital navigators around the country asking similar questions to help their clients. Could an AI tool be geared to answer the kinds of questions digital navigators get asked the most? And if creating that kind of AI tool could somehow be focused and specialized — then maybe it could be effective and scalable. And right nowas the funding for digital navigators has been pulled — some digital equity organizations are starting to lean into AI to assist in this area .

Nobody Is Regulating AI

Digital equity organizations need to be implementing AI usage based on their internal analysis — and that’s not just where AI fits into the work that they do. If they're not taking the time to understand how AI can be used in their work, they're already behind. And the people they serve are even further behind. They have to really figure out where in their work they're going to embrace AI and where they're not.

Some groups have major concerns about bias and how information is used against people of color. There are environmental organizations with concerns about data centers. There are advocacy organizations that may decide their role is to push for better restrictions. There are a lot of different ways AI could fit into the day-to-day work organizations do. But they have to acknowledge that it’s probably here to stay and have a plan for how they're going to engage around it.

AI feels unchecked because it is unchecked. You're not even allowed to check it at the state level. There's no safety net. If it were a drug, you would hope it would have been vetted better than it currently is.

The digital equity sector is asking who is ultimately responsible for closing the AI equity gap. Is it the government? The AI companies themselves? A third-party network of social impact professionals? Or is it all of them coming together? 

There are a lot of unknowns. But we should acknowledge the realities of what's coming, and the fact that nobody really knows how it's going to progress — because it's moving rapidly.

What AI Can Actually Do for the Communities We Serve

None of this means AI has nothing to offer the communities we work with. It does — for those who can access it safely and with proper support.

Translation is one of the clearest examples. Some of the translation tools have gotten really, really good, and they're making a practical difference for organizations serving multilingual populations.

AI can also help people write letters for specific situations — formatting them correctly, getting them in good shape, and not just translating but actually help writing something powerful. That's not an unusual need among the people served by digital equity organizations.

But all of it still comes down to digital literacy training. People always need training — to understand how to utilize AI, to help with writing prompts, to know how to stay safe, and how not to overexpose information.

The potential benefits of AI in digital equity are real. They just need to be supported by an ecosystem designed to bring everyone along.