As AI seeps into how we work, communicate, and make decisions, it’s quietly rewriting the terms of everyday life. Digging out patterns we’d never be able to notice—like the ones that can save lives, strengthening resilience and adaptation to climate-related hazards and swallowing grunt work that used to eat up our days. But it has also spawned an influx of texts, images, signals, videos, videos of humans morphing into vegetables (no, truly), fake videos of natural calamities that alarm residents and confuse first responders (which is mildly annoying at best and a costly nuisance at worst). And spam, so much spam. It can be so overwhelming that engaging with anything visual now means battling with emotions and finding ways to understand what information is real and what is imagined. 
In the past, information was often made more digestible and relevant through design choices–and was often thought of and used as the tool that makes things easier to understand.

Design has never been just window dressing. In fact, in a world battered by the onslaught of information and fractured attention spans, design’s original mandate can come to the fore: to cut through the noise and give the gift of clear communication–or in Busara’s case, clear communication of research. 

Good visual design ensures that people see, understand, and can engage with research findings.  It wrests meaning from the sea of information and plants it where it matters – in your head. When designing, we often take raw, unengaging, and convoluted data and present them in a manner that highlights what matters, signals what to focus on, and creates mental cues that guide understanding without requiring extra effort.
Behavioral science is only one place where design is used in this way. In finance, the framing of information affects how individuals perceive risk, make investment choices, and manage budgets. Clear visualizations of spending, debt, or investment performance can help people make informed decisions, while cluttered or misleading presentations can be detrimental.
In public health, clear, well-structured visuals can guide people toward beneficial behaviors—such as getting vaccinated, adopting hygiene practices, or following safety guidelines—while poorly presented data can cause confusion, fear, or inaction.

How visuals and AI work together at Busara

At Busara, graphics and visual elements are central to how we package and communicate our research outputs,  to illustrate findings, highlight insights, engage diverse audiences, and convey complex ideas with clarity. It is part of how we think, how we tell stories about the work we do, and how we make conclusions legible. 

AI is not only helping us do that–it also allows us to tackle a challenge that many organizations doing research with people have. We all want to show the people who are our research participants–but we also don’t want to endanger their privacy.

Specifically with photographs of people, standard practice has been to use real photographs of participants.  While that looks  authentic, in reality, it’s a bit of a minefield. Consent has to be sought (also for the image to be used in perpetuity). Even when consent is given, participants might not fully grasp where their images will travel, what they’ll represent, or how they’ll live on long after the project wraps.

And images do live on, sometimes for too long. The same photo, the same face, recycled across decks, reports, and presentations until it becomes a kind of fixture. What begins as representation quickly turns into repetition, or worse, misrepresentation. Beyond the fatigue, there’s risk, ethical and legal, considering autonomy, privacy, and confidentiality must be respected . We’ve seen it happen elsewhere, identifiable images originally meant for restricted academic use are broadly surfacing online, stripped of context and control. It’s not just uncomfortable, but also irresponsible.

We’re exploring how generative AI can help us move past this bind, not to sterilize our visuals, but to make them truer to context without borrowing anyone’s likeness. With the right tools, we can create images that reflect the diversity and dignity of our participants while preserving their privacy.
It’s about respect, yes, but also about control. About not having someone’s face resurface years later on a slide deck or report they never agreed to. For me personally, It’s the quiet  relief of being able to tell the story without turning anyone into an exhibit.

Moreover, here where co-design workshops drive much of what we do, AI lets us mock up visuals in real time. We can show participants a concept, get feedback immediately, and adjust on the spot. Traditionally, we’d sketch something, wait a day or two for review, then send it back for feedback, stretching what could be a single conversation into a drawn-out, multi-day process. With AI, the work moves at the pace of ideas, keeping iterations immediate and actionable.
The visuals below are draft posters and images generated by Malhar Acharya,  using DALL-E for a project on child marriage and female genital mutilation in Ethiopia. What’s wild is how quickly these came together. Generative AI has blown the doors off the creative process, anyone on the team (even non-visual design specialists)  can jump in, sketch out a thought, and see it take shape right there on the screen. That kind of speed changes the rhythm of the work, reduces lag time, and allows ideas to evolve on the fly.

Allowing ideas to take form quickly is crucial, it lets the team test concepts, gather feedback, and iterate before committing to final designs. These drafts are ready to be refined and evolved as the narrative strategy develops. And the visual design team at Busara, we are on the lookout for tools and technologies that boost creativity and have found a way to use AI in a way that aligns with that notion.

But the truth is, AI-generated visuals can be unrealistic, generic, or worse, skewed by bias. It does warrant discretion since AI-generated images also have a tendency to lack emotional depth and nuance. 

Which is why it is good to note that when using AI, every image it spits out is a direct reflection of the prompt you feed it, vague prompts yield vague images, while detailed, nuanced instructions produce something sharper and more purposeful.
This is why careful visual design is essential. At Busara, research findings inform the creation of personas, which then guide the prompts used to generate visuals. This approach ensures that the images reflect the target audience accurately, communicate insights effectively, and support rapid iteration and feedback during the design process, rather than producing misleading imagery.
The images below show just that process in action, from the information in the persona document on women-owned MSMEs (micro, small, and medium enterprises) in Kenya we were able to write a Midjourney prompt which generated an accurate visual representation of the reality of such an individual.

AI solves practical problems, it speeds up iteration, reduces reliance on real participants’ images, and makes visual outputs easier to work with.
But these conveniences don’t, or rather can’t, come without a cost. Training large generative AI models consumes staggering amounts of energy, leaving a carbon footprint that the average user barely registers. The datasets feeding these models can encode biases, sometimes subtle, sometimes glaring, reflecting historical inequities or skewed representations, and there’s no guarantee future outputs will be neutral or fair.
And this is just what we know for now, the long-term effects, social, environmental, and ethical, are still unfolding, quietly, in the background.

At Busara, we try to use AI deliberately and tactically, and my take on it is that generative models are powerful tools, they help solve problems, speed up co-design, and turn ideas into visuals almost instantly. At the same time, they raise important questions, consume energy, carry biases, and can produce results that need careful interpretation. The need of the hour is to use technology thoughtfully, understand its limits and guide how it is applied. 

In the future, behavioral scientists, designers, and AI will need to work together. It’s not just about making things look good, it’s about making them work better for people.
Visual design is a tool for influencing behavior, and for those in behavioral science, using visual storytelling in the appropriate way will be key to making interventions that are smart and effective.