Tailoring Your Message
Communicating AI Research Across Audiences
The Communication Paradox
The more you know, the harder it is to explain simply
The simpler the explanation, the more work it takes
Many of you mentioned struggling with public speaking or feeling ashamed of it. Here's the truth: communicating complex research across audiences is genuinely difficult. It's a skill separate from doing the research itself, and it takes practice.
But it's also increasingly necessary. Your work is "theory and research-based" while the public sees "a lot of information." The gap between your research reality and public understanding has real consequences—for funding, for policy, for public trust in AI.
You don't need to be a professional communicator. But you do need to be thoughtful about how you translate your work for different audiences. Consider it part of your professional responsibility, not an annoying add-on.
The Communication Challenge
Today we're tackling something that came up repeatedly in your intake forms: the challenge of communicating the same research to vastly different audiences. Whether you're talking to fellow researchers, industry partners, policymakers, or the general public, your core research stays the same—but your approach must change dramatically.
Your Lab Collaborators
Shared technical language and deep domain expertise
Industry Leaders
Focus on practical applications and business value
Policy Makers
Concerned with societal impact and regulation
The Lay Public
Need accessible explanations without jargon
You're essentially operating in four different communication ecosystems simultaneously. Each has its own language, priorities, and expectations. The mistake many researchers make is trying to use one approach for all four. That's like trying to use the same key for four different locks.
Know Your Audience Framework
Before we dive into each audience type, let's establish a framework. For every communication situation, ask yourself these five questions. This systematic approach will help you tailor your message effectively and avoid common communication pitfalls.
01
WHO
Who are you talking to specifically? Not just "industry" but which industry stakeholders? What's their background and expertise level?
02
WHAT
What do they already know? What do they need to know? What do they think they know that might be wrong?
03
WHEN
When are you communicating—before funding decisions? After a breakthrough? During a crisis? Timing shapes context.
04
WHERE
Where will this communication happen—formal presentation, casual conversation, written report? Medium matters.
05
WHY
Why does this audience care about your work? This is crucial. Your "why" will be completely different for each group.
In a corporate context, these are often called "personas", you can find lots of resources and thinking about how to create these assets that might be useful.
Communicating with Fellow Researchers
Shared Language, Nuanced Challenges
Several of you mentioned that your most common audience is "other researchers" or "scientists from other fields." This might seem like the easiest audience, but it has unique challenges.
The Curse of Knowledge or the collaborative challenge
You know so much that it's hard to remember what needs explaining. One intake response mentioned the difficulty of "starting at the right point so nothing is left out." This is the collaborator's dilemma.
Lead with the Gap
Start by addressing the gap in existing research you're tackling
Explain Established Frameworks
Leverage your field's terminology and conceptual models but make sure to explain them
Focus on Methodology
Emphasize reproducibility and technical rigor
Be Specific About Limitations
Fellow researchers will spot them anyway—address them proactively
Frame Your Contribution
Position your work within ongoing conversations in your subfield
Set yourself in history
You collaborators likely don't havre your long term memory, who are your heroes and why?

Example Structure: "Current approaches to [X] face challenges with [Y]. Our method addresses this by [Z]. We validated this on [dataset] and found [results], though limitations include [constraints]."
Cross-Disciplinary Collaborators
Building Bridges Between Fields
This is where it gets interesting. Working in project consortia with researchers from other disciplines presents unique challenges. Medical researchers think in terms of clinical outcomes; you think in terms of model performance. These aren't always aligned.
The Translation Challenge
You're both experts, but in different domains. The key is finding shared conceptual ground and building from there.
  • Find the shared conceptual ground first
  • Use analogies from their field: "This is similar to how in medicine you..."
  • Define your technical terms once, clearly
  • Focus on implications for their domain
  • Be explicit about what the AI can and cannot do
Red Flag:
Managing expectations about AI's capabilities is crucial. Your collaborators in other fields may have inflated expectations from media coverage.
Communicating with Industry Leaders
From Research to Results
Industry audiences care about different things than academic audiences. They're asking: "Can this work at scale? What's the ROI (return on investment)? What are the risks?" Start with the business case, not the technical innovation. Then work backward to why your research makes it possible.
Practical Application
Can this be deployed in real-world scenarios? What's the path to implementation? Are there ethical considerations?
Competitive Advantage
How does this create market differentiation or improve existing solutions?
Risk and Liability
What could go wrong? What are the failure modes and mitigation strategies?
Timeline and Resources
What investment is required and when will we see returns?
Market Impact
How does this affect our position in the marketplace?
Ask questions
These people are experts as well, but the best sign of respect is interest and engagement
Industry Communication Structure
The Opportunity
"Current solutions cost X or fail at Y rate"
Your Innovation
"Our approach reduces this by Z%"
The Validation
"We tested this on [real-world scenario]"
The Path Forward
"To deploy this requires [resources/timeline]"
The Risks
"Current limitations include..."
What They're Really Asking
  • "How is this better than what we're already doing?"
  • "What will this cost to implement?"
  • "What could go wrong?"
  • "Who else is doing this?"
  • "How do I explain this to my stakeholders?"
Secondary Audiences
Notice that last question—industry leaders also have audiences. Your research might go to their board, their investors, their clients. Make your explanation packageable for them to pass along.
Policy Communication Approach
01
Frame Around Their Concerns
Don't lead with your research; lead with the societal issue they're grappling with
02
Explain Implications, Not Mechanisms
They don't need to understand backpropagation; they need to understand what reliable vs. unreliable AI means for public policy
03
Provide Actionable Insights
"Here's what we know, what we don't know, and what that means for regulation"
04
Use Concrete Examples
Abstract technical explanations won't land; specific scenarios will
05
Address Fears Directly
Policymakers hear concerns about job replacement and AI takeover from constituents—acknowledge these
Three-Part Framework for Policy Communication
BEGINNING: Current State
  • "Right now, AI systems in [domain] can/cannot..."
  • "The public is concerned about..."
  • "Existing approaches face challenges with..."
MIDDLE: What's Changing
  • "Research shows that..."
  • "Our work demonstrates..."
  • "This means that [specific capability/limitation]..."
END: Implications
  • "For policymakers, this suggests..."
  • "Regulatory frameworks should consider..."
  • "We don't yet know [X], which means..."

Notice what's NOT in this structure: technical methodology, mathematical proofs, model architecture. Those details don't serve the policymaker's needs. Always acknowledge uncertainty—policymakers actually appreciate this more than false certainty.
Communicating with Policy Makers
Translating Uncertainty into Action
This is perhaps the most challenging audience. Policymakers need to understand the implications of uncertainty and reliability in AI, but don't need full mathematical derivations. They're high-stakes decision makers with limited technical expertise.
Public Safety and Welfare
How does this technology affect citizens' wellbeing and security?
Economic Impact
What are the broader economic implications for jobs, industries, and growth?
Regulatory Frameworks
What rules and oversight mechanisms are needed?
International Competition
How does this position us relative to other nations?
Constituent Concerns
What are voters worried about and how do we address those fears?
Your Biggest Challenge: Helping people feel comfortable with "We don't know, here's why, and here's how we quantify that uncertainty" takes time and trust. This is fundamentally different from the certainty policymakers often seek.
Communicating with the Public
Overcoming the Enchantment
Here's where we need to confront something fundamental: the word "intelligence" changes the context without us realizing it. Intelligence is "at the heart of what we believe makes humans special" and forces us into anthropomorphizing.
Common Public Misconceptions
  • "AI thinks like humans"
  • "AI can replace jobs/humans/everything"
  • "AI is smarter than humans"
  • "LLMs are close to general AI"
Your Goal
Move the conversation toward everyday/realistic applications while managing both utopian hype AND dystopian fear. When you say "artificial intelligence," you've already primed people to think in anthropomorphic terms.
Public Communication Strategy
Start Where They Are
1
Assess Their Starting Point
What do they already think about AI? (Probably something they saw in media)
2
Identify Their Frame of Reference
Movies? ChatGPT? Alexa? Smart home devices?
3
Connect to Their Lives
What do they care about in their daily experiences?
What to Avoid
  • Don't use anthropomorphic language ("the AI thinks," "it knows," "it decides")
  • Don't start with what your research CAN'T do (leads with negative)
  • Don't overwhelm with caveats (one caveat is enough)
  • Don't assume mathematical knowledge
What Works
  • Concrete, relatable examples
  • Visual demonstrations
  • Analogies to things they know
  • Focus on specific applications, not general capabilities
  • Your genuine enthusiasm (but tempered with honesty)

The Metaphor Approach: "Think of it like..." is your friend here. Gameplay is like reinforcement learning—that's a great metaphor for lay audiences. Use familiar concepts to bridge to unfamiliar technology.
The Story Structure for Public Audiences
1
Once Upon a Time
"Doctors looking at MRI scans..."
2
Every Day
"...spend hours analyzing each one..."
3
One Day
"...what if AI could help identify patterns faster?"
4
Because of That
"...we developed a method that..."
5
Until Finally
"...doctors can focus on treatment rather than detection"
This structure works because it's how humans have told stories for millennia. It doesn't feel like a technical presentation; it feels like a narrative they can follow. Notice this example didn't mention neural networks, training data, or model architecture. For lay audiences, those are obstacles to understanding, not aids.
"The best explanations don't simplify the complexity—they make the complexity accessible through narrative."
Common Ground Across All Audiences
Universal Principles
Regardless of your audience, these principles apply. The execution changes, but the foundation doesn't. Every audience respects appropriate humility more than false confidence—the key is appropriate.
Know What They Care About
It's not always what you care about—understand their priorities and motivations
Start Where They Are
Not where you are—meet them at their level of understanding
Choose Your Vulnerability
What you admit you don't know builds trust and credibility
One Clear Message
One clear message is better than three muddled ones—focus and simplify
Show, Don't Just Tell
Demonstrations and examples are more powerful than abstract explanations
On Vulnerability: Appropriate for Each Audience
Collaborators
Be specific about technical limitations
Industry
Be clear about deployment constraints
Policymakers
Be honest about unknowns that affect decisions
Public
Be straightforward about what AI can't do

One Clear Message: If your audience remembers only one thing, what should it be? Build everything else around that central insight.
Practice Exercise
Your Research in Four Sentences
Write one sentence for each audience that captures your work's essence for THEM. You'll quickly see how different these need to be. And that's okay—that's the point.
1
For Collaborators
Technical language, methodology focus, specific contribution to field
Example: _______________________________________
2
For Industry
Business value, practical application, measurable impact
Example: _______________________________________
3
For Policymakers
Societal implications, regulatory considerations, actionable insights
Example: _______________________________________
4
For Public
Relatable examples, clear benefits, accessible language
Example: _______________________________________
Resources & Next Steps
For Day 2 Projects
Different audiences aren't obstacles—they're opportunities to make your research matter in different ways. For your Day 2 presentations, you'll apply these principles to create a communication piece in your chosen medium for a specific audience.
1
Choose Your Target Audience
Be specific about who you're trying to reach
2
Define Their "Why"
Why should they care about your work?
3
Select Your Medium
Choose the format that best serves your audience and message
4
Structure Your Message
Use frameworks from today to organize your content
5
Practice Vulnerability
Be honest about limitations in ways appropriate to your audience

Think carefully about who you're trying to reach and why they should care. The technical brilliance of your research remains the same; how you communicate it must adapt.
Moving Forward
Building Your Communication Toolkit
Communication excellence doesn't happen overnight. It's a skill you develop through practice, reflection, and iteration. Each presentation, each conversation, each written piece is an opportunity to refine your approach.
Death to frankendeck!
Seek opportunities to present to different audiences. Each experience builds your adaptability and confidence. Make sure each talk or presentation is tailored to your current audience, not your last one
Seek Feedback
Ask your audiences what landed and what didn't. Their insights are invaluable for improvement.
Reflect and Iterate
After each communication event, note what worked and what you'd change. Build your personal playbook.
Embrace Growth
Communication skills improve with deliberate practice. Be patient with yourself as you develop this crucial ability.
The gap between research reality and public understanding has real consequences. Your ability to bridge that gap makes your work more impactful, more fundable, and more likely to create positive change in the world.