Workers everywhere are embracing artificial intelligence (AI) in different ways to enhance their work: Gallup data shows that from 2023 to 2025, the percentage of U.S. employees who say they have used AI in their role a few times a year or more nearly doubled, from 21 percent to 40 percent. It’s likely this has shot up even further this year.
This has sparked the imaginations of those of us designing eLearning, with course designers dreaming of a much faster process, more personalized content, and realistic role-play simulations using conversational agents. Can the administrative burden be exponentially lessened, while content quality exponentially increases?
The answer is, broadly, “Yes,” but for those of us in more risk-averse sectors, enthusiasm alone isn’t enough, and caution still rules. For course designers working in financial services, law, and beyond, we’re stuck in a paradox: Evidence is required before adoption, yet adoption is the only way to generate evidence.
The good news is that you don’t have to start with anything learner‑facing or confidential. You can start in the quiet corners of the workflow—the parts that are high-effort, low-glamour, and often neglected. Here are some practical, low-risk ways to test, learn, and build confidence, all while staying firmly within compliance guardrails.
A New Golden Age for Accessible eLearning
One of the primary arenas for this is a part of eLearning close to my heart, but it can be linked with significant time and energy demands: making learning more accessible.
Not long ago, this demanded hours of meticulous effort. Writing decent alt text, checking color contrast, simplifying dense passages without losing meaning, and evaluating your design choices to make sure they don’t accidentally exclude people.
This is where AI can help, not by doing it for you, but by giving you a strong first draft so you’re not starting from zero.
For example:
- If a paragraph is overly complex, ask an AI tool to rewrite it in plain English (then you decide what stays or what changes).
- If you have an image, you can use AI to generate a first pass at alt text, then refine it with your own judgment.
- If you’re working with a fixed brand palette, you can use AI to suggest accessible combinations and placements, applying that essential human check where it matters (color contrast is still one area where I definitely wouldn’t fully outsource the final call).
There is so much value in the momentum AI can create, so accessibility doesn’t become the thing that is pushed to “later.”
Multimedia Made Easier
Meanwhile, tools such as Adobe Premiere Pro now produce transcripts and closed captions with impressive reliability. You also can streamline the correction process and prompt the AI to highlight likely errors that need reviewing. This is also a great step to add to your final quality check process to catch any small errors that could fall through the cracks.
Importantly, for this kind of experimentation, you must be careful not to share confidential information with your chosen AI tool. It’s just about enhancing accessibility and inclusion, so redacting the information shared also may be an important step. If you’re fortunate enough to have access to a secure, internally managed large language model (LLM), it can safely review sensitive or intellectual property-protected material without you needing to worry about that extra step.
Again, the point isn’t to pretend it’s flawless; it’s to reduce friction and give learning designers more bandwidth to focus on the learner experience and the quality of practice. Used thoughtfully, AI becomes less about risk and more about removing barriers; helping the right information reach more people, without undermining security.
Translation and Localization Support
AI also can produce strong first-draft translations and suggest cultural adaptations, significantly reducing the overall workload.
Specifically, early AI intervention can prevent costly rewrites by reviewing pre-translated content for idioms, colloquialisms, or culturally specific references that are unlikely to travel well, allowing content teams to resolve potential issues before formal translation begins.
That said, not all LLMs are created equal across languages. Performance for translating specific languages can vary from one AI tool to another (sometimes dramatically), so careful selection matters. And however capable the output appears, bypassing human review simply isn’t an option—AI can accelerate the process, but it cannot assume responsibility for cultural accuracy.
AI as a Creative Sidekick
Perhaps most obvious is AI’s use in creative idea generation. You can quickly draft scenarios or mock up visual concepts shaped around a defined audience, tone, or set of learning objectives.
It doesn’t replace seasoned designers or learning specialists, nor should it, but for smaller teams or projects constrained by time and budget, it can act as a force multiplier, helping ideas take shape faster. Sometimes, just a small shift in angle, a prompt that reframes the problem, or a structure you hadn’t considered can be the catalyst that gets a project across the finish line.
Whether drafting skeleton outlines, proposing alternative learning objectives, or suggesting new ways to chunk content, the key to keeping things low risk here is providing clear guardrails within your AI prompts to define the boundaries of its contribution. You are being careful to ensure no confidential material is shared, so the AI tool can be used as a sounding board for structure and perspective, like testing ideas with a colleague. AI can accelerate the thinking, while human judgment remains the final authority.
All of the above allow those of us in tightly regulated industries to finally amplify our expertise with automation. AI adoption doesn’t have to begin with high-stakes decisions or learner-facing risk. It can start in these quiet, practical corners of our workflow, providing low-risk, high-impact use cases that build familiarity, evidence, and internal confidence. Each small win chips away at that paradox.
The golden thread running through everything is judgment. Handled carelessly, AI introduces risk; handled thoughtfully, it removes friction, freeing designers to focus more on crafting meaningful and inclusive learning experiences.
