How AI is Changing the LXD Role: What to Embrace, What to Rethink, and What Will Always Be Yours

Let me say this plainly, because too many conversations about AI and L&D are still dancing around it: AI is not just another tool in your toolkit. It is not like the transition from Word to Google Docs, or from PowerPoint to Articulate. It is a structural shift in what Learning Experience Designers are paid to do, how long it takes to do it, and — most importantly — which parts of our work actually require a human being. That is an uncomfortable thing to sit with. It is also the most important thing you can be honest about right now if you want to navigate the next five years with clarity rather than anxiety.

I have been doing this work for over a decade. I have lived through the “but will rapid eLearning kill quality?” debate, the “will microlearning replace courses?” panic, and about three cycles of “AI is going to change everything.” This time, though, I am not going to tell you that nothing fundamental is changing. It is. And if you are a senior LXD who has built your career on being good at producing things — writing scripts, building modules, creating storyboards — you need to read this clearly and think hard about what comes next.

This article is my honest attempt to map what AI is actually automating, what it cannot touch, and what that means for how you position yourself and evolve your practice.

What AI is Actually Automating — Let’s Be Specific

I am tired of vague statements like “AI will assist with content creation.” Let’s get specific about what is happening right now, in 2025, in real organizations. These are not hypothetical capabilities. These are things happening today with tools that cost less than a Netflix subscription.

First-Draft Course Content and Scripts

ChatGPT, Claude, and Gemini can generate a first-draft eLearning script for a 15-minute module in under two minutes, given a topic, audience description, and learning objectives. Is the output always good? No. Is it a passable foundation that an experienced LXD can edit in an hour rather than write from scratch over a day? Increasingly, yes. The task of “writing the first draft” has been radically devalued. Not eliminated — but devalued.

Storyboarding and Module Structure

Tools like ChatGPT with structured prompting can generate a complete module outline — with section breakdowns, learning activity types, estimated times, and knowledge check placement — faster than you can open a new document. Articulate AI within Rise 360 now offers content block suggestions as you build. The cognitive work of “how should I organize this module?” is being partially offloaded.

Quiz and Knowledge Check Generation

This is perhaps the most fully automated of all LXD tasks. Given learning objectives and source content, every major AI platform can generate banks of multiple-choice, true/false, scenario-based, and short-answer questions in seconds. Questions that would have taken a junior designer an afternoon now take thirty seconds. The difference between good and bad AI-generated questions is substantial — but the production time is gone.

Image and Visual Asset Creation

Midjourney, DALL-E 3, and Adobe Firefly generate custom learning illustrations — scenario characters, conceptual visuals, infographic elements — on demand. Organizations that used to license stock photography or pay for custom illustration are rapidly shifting. The art direction still matters enormously. The hours of searching, licensing, and resizing do not.

Video Narration and Voiceover

ElevenLabs and Murf produce near-human-quality synthetic voiceover. Synthesia and HeyGen generate full AI avatar video with lip-sync narration. Organizations that could never afford professional voiceover talent are now producing narrated video at scale. Organizations that could afford it are rethinking the cost. The market for “record the narration” as a task has fundamentally changed.

Translation and Localization

AI translation — particularly through DeepL and GPT-4o — has made first-pass localization of eLearning content practical at a cost and speed that human translation cannot match. Human review for accuracy and cultural nuance remains essential. The “translate this 45-minute course into Spanish, French, and Portuguese” project that used to take weeks and significant budget now takes hours and a fraction of the cost.

Basic eLearning Module Scaffolding

Articulate AI, Adobe Captivate AI, and emerging tools like Coursebox can take a PDF or document and generate a structured eLearning module — with slides, content blocks, and basic interactions — from uploaded source material. It will not be good without a human shaping it. But the empty-canvas problem is largely solved by AI.

Summarizing SME Interviews and Source Documents

Upload a recorded SME interview transcript to Claude or ChatGPT and ask it to extract the key learning points, identify gaps, suggest follow-up questions, and organize insights by topic. What used to take an ID two to three hours of careful reading and synthesis now takes ten minutes. The relationship with the SME, the ability to know what questions to ask, the judgment about what matters — those remain human. The administrative synthesis work is largely automatable.

Writing Learning Objectives from Content

Generate a content document, specify Bloom’s taxonomy level, and AI will produce a set of learning objectives faster than you can think of them. Are they always perfectly calibrated? No. Do they require human review and refinement? Yes. But the starting point is no longer a blank page.

The Skills That Are Becoming More Valuable

Here is the thing about structural shifts: they do not erase human expertise, they reallocate it. When spreadsheets arrived, the skill of manual calculation became less valuable. The skill of knowing what to calculate and why became more valuable. AI is doing the same thing to learning design. The production tasks that consumed enormous time are contracting in value. The judgment, strategy, and relationship skills that underpin those tasks are expanding in value.

Performance Consulting — Before You Design Anything

This has always been the highest-value LXD skill, and AI is making it even more differentiating. Performance consulting means diagnosing the actual root cause of a performance gap before defaulting to a training solution. It means asking: is this a skill problem, a motivation problem, a process problem, an environment problem, or a combination? It means having the confidence to walk into a stakeholder meeting and say, “I don’t think training is your answer here.”

AI cannot do this. AI can generate training content for any topic you give it. It cannot walk into an organization, read the dynamics, have conversations with frontline workers and senior leaders, and diagnose why performance is actually falling short. That consulting skill — grounded in models like Gilbert’s Behavior Engineering Model, the Performance Consulting framework, and EPSS thinking — is becoming the core differentiator for senior LXDs.

Learning Strategy and Program Architecture

AI generates content. It does not design learning ecosystems. Learning strategy — deciding what combination of modalities, touchpoints, reinforcement strategies, social learning, and performance support will actually move the needle for an organization — requires deep understanding of adult learning theory, organizational context, and the specific performance outcomes at stake. This is big-picture work. It requires synthesis across business strategy, L&D capability, technology constraints, and culture. AI cannot do it. Senior LXDs who have been doing it intuitively need to make it visible and explicit as a competency.

Stakeholder Management and Influence

The ability to work with senior leadership, translate business needs into learning strategy, manage scope, push back on bad ideas diplomatically, and build credibility over time is fundamentally relational. AI cannot manage your relationship with the VP of Sales who wants a two-day training course when what the team actually needs is better onboarding materials and a coaching cadence. That is human work.

Instructional Quality Judgment

This is underrated and will become one of the most important skills in the field: the ability to look at AI-generated learning content and know when it is wrong. Not just factually wrong — though that happens constantly with AI hallucinations — but pedagogically flat. Cognitively passive. Structurally off. Motivationally inert.

AI-generated content is often technically accurate and completely uninspiring. It frequently produces explanations that are clear but do not create the cognitive challenge needed for durable learning. It creates scenarios that are procedurally correct but emotionally unconvincing. Knowing the difference — and being able to fix it — requires deep expertise in how learning actually works. That expertise is not in the AI. It is in you.

Facilitation and Human Learning Dynamics

Live learning — workshops, coaching conversations, cohort programs, communities of practice — is experiencing a quiet renaissance precisely because so much content delivery has moved to self-paced AI-generated modules. The skill of facilitating genuine human learning, creating psychological safety, managing group dynamics, and building the conditions for insight and behavior change is irreplaceable. If you have neglected facilitation skills in favor of production skills, reconsider.

Ethical Oversight and AI Governance

Organizations are generating learning content with AI faster than they are thinking about whether they should. Someone needs to be the adult in the room who asks: Is this accurate? Is this biased? Is this appropriate for the audience? Does it comply with regulatory requirements? Is the learner informed that this was AI-generated? Is this interaction designed by AI something we are comfortable putting in front of our employees?

That person is increasingly the LXD. AI governance in learning is a genuine emerging responsibility, and the LXDs who develop expertise in it will be indispensable.

Data Literacy and Learning Analytics

xAPI, learning record stores, completion rates, assessment performance, behavior change metrics — the organizational data around learning is richer than it has ever been, and most organizations are doing almost nothing with it. LXDs who can read this data, draw conclusions about learning effectiveness, and make evidence-based recommendations for program improvement have a significant advantage. Watershed, Learning Locker, and similar platforms produce data that most L&D teams are not equipped to interpret. Be the one who can.

Prompt Engineering as a Core Professional Skill

Let’s be direct: in 2025, prompt engineering is not a nice-to-have for LXDs. It is a professional competency. How you prompt AI determines the quality of what you get. An LXD who knows how to give AI detailed, structured, pedagogically-informed prompts will get dramatically better outputs than one who types “write me a course about customer service.” Learning to prompt well — with context, constraints, persona, format specifications, and iteration — is something you should be practicing daily.

The New LXD Workflow: Before AI vs. After AI

Let me paint this picture honestly, because abstract descriptions of “AI-augmented workflows” can sound much more comfortable than the reality.

Before AI (2020 and Earlier)

A mid-size eLearning project — let’s say a 30-minute compliance module — involved: stakeholder interviews (2-4 hours), SME meetings and follow-ups (4-8 hours), content research and synthesis (4-6 hours), writing the script (8-12 hours), building the storyboard (4-6 hours), development in Storyline or Rise (12-20 hours), review cycles (4-8 hours), voiceover coordination (2-4 hours), QA and finalization (3-5 hours). Total: roughly 40-70 hours of designer time, often spread over 4-6 weeks.

After AI (2025)

The same project: stakeholder conversations (still 2-4 hours, non-negotiable), AI-assisted content synthesis from SME transcripts (30-60 minutes instead of hours), AI first-draft script with your structured prompt (20-30 minutes, plus 1-2 hours of refinement), module structure generated and refined in Rise with AI block suggestions (2-3 hours), AI-generated visuals through Adobe Firefly or Midjourney (1-2 hours instead of stock hunting), synthetic voiceover via ElevenLabs (30 minutes instead of days), AI-generated knowledge checks refined by you (30 minutes), review cycles (still 4-8 hours), QA and finalization (2-3 hours). Total: roughly 15-25 hours, often compressed into 2-3 weeks.

The “AI as Junior Designer” Mental Model

The mental model I find most useful — and most honest — is this: treat AI as a very fast, very tireless, moderately skilled junior designer. It needs clear direction. It makes mistakes you have to catch. It produces work that is good-but-not-great without your refinement. It cannot make strategic judgments. It does not understand your organization, your stakeholders, or the specific sensitivities of your content. You are still the lead designer. AI is doing the execution work so you can focus on the direction work.

This mental model breaks down when organizations start expecting LXDs to manage five times as many projects simultaneously because “AI makes everything faster.” That is a conversation we need to have with leadership explicitly: more projects is not the right use of time saved. Better learning is.

The Quality Question

More content does not mean better learning. This is the hill I will die on. The risk of AI-accelerated L&D is not that the content will be bad — it is that organizations will interpret faster production as license to produce more content rather than better learning experiences. A poorly designed 30-minute module produced in 15 hours is no better than one produced in 60 hours. The design thinking is what matters. Do not let time savings get consumed by volume inflation.

Roles That Are Evolving — And How

Instructional Designer → Learning Experience Architect

The traditional Instructional Designer role was heavily weighted toward production: write content, build modules, create assessments, develop storyboards. That work is not disappearing — but it is being compressed by AI in ways that change the value equation. The evolutionary path leads toward Learning Experience Architect: someone who designs the overall learning journey, makes decisions about modality and sequencing, frames the learner experience from start to finish, and evaluates whether the resulting programs actually change behavior. Less execution, more design thinking. Less module building, more ecosystem mapping.

eLearning Developer → AI-Augmented Developer

eLearning developers who focused on Storyline programming, interaction design, and technical production are finding that AI handles much of the boilerplate. The developers who are thriving are those redirecting their time toward complex interactions, custom JavaScript, accessibility optimization, and the technical integration work that AI cannot do reliably. They are using AI to scaffold, and their expertise to finish. Those who resist this shift and continue to position themselves purely on production speed will face pressure.

L&D Manager → Learning Technology Lead

The L&D Manager role in 2025 increasingly involves evaluating, selecting, and governing an AI-augmented learning technology stack. Which content generation tools do we use? What are our governance policies for AI-generated learning? How do we validate AI outputs? What does our xAPI data strategy look like? These are not purely technical questions — they require someone who understands both learning and technology deeply enough to make good decisions. The L&D Manager who positions as a learning technologist and strategic advisor to the business will have expanding influence.

Content Developer → Content Curator and Quality Lead

The Content Developer role — focused on writing and organizing instructional content — is the role most directly disrupted by AI content generation. The evolutionary path is toward Content Curator and Quality Lead: someone who evaluates AI-generated content for accuracy, pedagogical soundness, organizational fit, and audience appropriateness. This is a real and valuable role. It requires deep subject matter knowledge and strong editorial judgment. It is, however, a role that can be done with less headcount than traditional content development teams. Organizations will adjust accordingly.

The Automation Anxiety Question

Let me address this directly, because I think a lot of the conversation in L&D communities is either catastrophizing or dismissing, and neither helps.

Is AI going to replace LXDs? Not as a field. But it will eliminate some roles and compress others. The honest breakdown:

The LXDs most at risk are those whose value is primarily transactional — getting content from source to screen efficiently. The LXDs who will thrive are those whose value is consultative — helping organizations understand what kind of learning they actually need and why, and whether any of it should even be a course.

What is truly irreplaceable: Judgment. Relationships. Organizational knowledge. The ability to read a political dynamic and know when to push and when to wait. The capacity to sit with a frustrated learner and understand what is actually blocking their performance. The experience to recognize when a training request is really a management problem in disguise. AI cannot replicate any of this. These are the things worth doubling down on.

How to Position Yourself in an AI-Augmented Field

Build a “Human + AI” Portfolio

Your portfolio in 2025 cannot just be a gallery of finished modules. It needs to demonstrate judgment. Show the before-and-after of AI-generated content you refined. Explain the strategic decisions you made about program architecture. Include the performance consulting work where you redirected a training request into a more appropriate intervention. Make visible the thinking that AI cannot replace. Any AI can produce a module. Show that you produce learning strategy.

Develop Visible Expertise in a Specialty

The generalist LXD who is good at everything faces more AI pressure than the specialist who is the go-to person for learning in a specific domain. Healthcare learning, leadership development, sales enablement, technical skills training, compliance learning — each has specific nuance that requires domain knowledge and context that AI does not inherently have. Develop and make visible your depth in a domain. Be the person organizations call because you understand their specific context, not just instructional design in the abstract.

Learn to Evaluate and Govern AI Tools — Not Just Use Them

There is a difference between being an AI user and being an AI-informed learning designer. The latter can evaluate tools against pedagogical criteria, identify risks and failure modes, establish organizational policies for AI use in learning, and make procurement recommendations grounded in learning quality standards. This is a genuinely emerging expertise and almost nobody has it yet. If you develop it, you will be the person your organization turns to when leadership says “we want to use AI for training” and somebody needs to ensure that actually results in learning.

Advocate for Learning Quality

Someone needs to be the organizational conscience when leadership gets excited about AI’s ability to generate training content in minutes. That person should be you. Advocate for the evidence base. Ask the evaluation questions. Challenge the “more content = more learning” assumption. Be the one who insists that faster production is only valuable if the learning design underlying the content is sound. This advocacy role is uncomfortable and important and increasingly rare.

Develop Business and Consulting Skills

If you have not invested in business acumen and consulting skills — how organizations make decisions, what drives executive priorities, how to present a business case, how to quantify L&D impact in financial terms — now is the time. The LXD who can have a strategic conversation with a CFO about the ROI of a learning program has a very different value proposition than one who can only discuss instructional design principles. Both matter. Only one will get you a seat at the table.

AI Tools Every LXD Should Know in 2025

This is not a comprehensive buyer’s guide — it is a practical orientation to the landscape as it stands, with honest notes about where each category of tools is genuinely useful versus overhyped.

Content Generation

• ChatGPT (GPT-4o): The most widely used, excellent for script writing, scenario generation, learning objective drafting, and general-purpose content. Tends toward verbose output that needs tightening. Strong on breadth, sometimes shallow on depth.
• Claude (Anthropic): My personal preference for instructional writing. Produces more nuanced, less templated prose. Better at following complex, multi-part formatting and pedagogical instructions. Excellent for long-form content requiring careful structuring.
• Gemini (Google): Strong integration with Google Workspace, which matters enormously in organizations running on Google tools. Best used when your content ecosystem lives in Google Docs and Slides.

The honest advice: learn to prompt all three well, and use them for different things. Do not become a single-tool person in a multi-tool world.

Visual Design

• Midjourney: Still the leader for creative, stylized illustration. Learning to prompt Midjourney for consistent learning characters across a module is a genuine skill. The results, done well, are remarkable.
• DALL-E 3 (via ChatGPT): More accessible, integrates directly into ChatGPT workflows. Good for quick conceptual visuals and realistic photographs. Less artistically consistent than Midjourney across a project.
• Adobe Firefly: The right choice for organizations already in the Adobe ecosystem. Strong for editing, compositing, and maintaining brand consistency. Trained on licensed content, which matters for enterprise copyright concerns.

Video

• Synthesia: The most mature AI avatar video platform. Hundreds of avatar options, strong lip-sync, multilingual. Output quality is good enough for most corporate learning contexts. Does not feel fully human, but learners have largely adapted.
• HeyGen: Stronger for personalized video at scale and for cloning your own likeness for video narration. Growing rapidly in capabilities.
• Descript: Slightly different use case — primarily for editing real video and audio, with AI features for removing filler words, overdubbing audio corrections, and transcript-based editing. Essential for anyone producing recorded learning content.

Voice and Audio

• ElevenLabs: The highest-quality AI voice generation available. The difference between ElevenLabs and older TTS tools is significant. For English-language narration in particular, output is often indistinguishable from professional voiceover. Multiple languages, voice cloning capability.
• Murf: More accessible, good library of pre-built voices, integrates with some eLearning tools. Not quite ElevenLabs quality but a strong option for organizations wanting simplicity.

Course Authoring with AI

• Articulate AI (Rise 360): AI content block suggestions, image generation, translation features integrated directly into the authoring tool. The integration reduces friction significantly for Rise-based workflows.
• Adobe Captivate AI: AI-assisted slide creation, text-to-course generation, responsive design assistance. Better for complex, interactive content than Rise but steeper learning curve.

Analytics and Insights

• Watershed: The most sophisticated xAPI analytics platform for learning. If your organization is serious about measuring learning impact beyond completion rates, this is the direction.
• Learning Locker (xAPI): Open-source LRS option for organizations with technical capacity to implement it. The foundation for any serious learning data strategy.

Workflow Automation

• Zapier + AI: Connect your content generation tools, LMS, and productivity apps. Automate notifications, reporting workflows, content routing for review. Not glamorous, but workflow automation saves real time.
• Make (Integromat): More flexible than Zapier for complex multi-step automations. Better option if you have a technically-oriented L&D team building sophisticated content pipelines.

The Ethical Dimension — This Is Not Optional

I want to spend real time here, because I see too many L&D teams treating AI ethics as a “we’ll figure it out later” problem. Later is now.

Transparency With Learners

Should learners know when the learning content they are consuming was generated by AI? I believe yes, with nuance. Learners have a reasonable interest in knowing whether the scenario characters they are responding to were designed by humans with their context in mind, or generated by a language model. The emerging expectation in many industries — particularly healthcare and finance — is disclosure. Get ahead of this with a clear organizational policy.

Accuracy and Hallucination Risk

AI language models hallucinate. They state incorrect information with complete confidence. This is a fundamental characteristic of how they work, not a bug that will eventually be fixed. In instructional content — particularly compliance training, medical education, safety training, or any domain where incorrect information has real-world consequences — every AI-generated claim requires human verification against authoritative sources. This is not optional. Building a verification workflow into your AI content process is a professional requirement.

Bias in AI-Generated Characters and Scenarios

AI image generators and text generators reflect the biases of their training data. AI-generated scenario characters skew toward certain demographics. AI-generated workplace scenarios reflect certain cultural assumptions. AI-generated names for fictional employees reflect biases in frequency of names across cultures. None of this is neutral. Reviewing AI-generated learning content for representation, cultural appropriateness, and demographic diversity requires conscious intention and human oversight.

Copyright and Intellectual Property

The legal landscape around AI-generated content ownership is still evolving. What we know: content generated by AI using training data you did not license may create intellectual property complications. Adobe Firefly explicitly trains on licensed content and indemnifies commercial use — which is why it matters for enterprise contexts. Other image generators carry more ambiguous IP risk. Organizations need a policy, and LXDs need to understand the basics well enough to advise on it.

Over-Automation of Learning That Should Be Human

This is perhaps the subtlest ethical issue: knowing when AI-generated, self-paced content is the wrong solution regardless of how efficiently it can be produced. Leadership development. Difficult conversations. Ethical reasoning. Grief and trauma support in healthcare. These are domains where the relational, human quality of learning is not a nice-to-have — it is the mechanism of change. AI-generated modules on these topics are not just lower quality; they are actively wrong for the job. Part of the LXD’s ethical responsibility is to name this clearly when organizational pressure toward automation threatens to commoditize learning that should remain fundamentally human.

What Will Always Be Yours

Let me make the closing argument clearly, because this is what I actually believe after a decade in this field and eighteen months of deep engagement with AI tools.

Building Trust With Learners and Stakeholders

Trust is built through relationships, consistency, credibility, and genuine care. A learner who knows their LXD listened to their feedback and shaped the program around their actual experience will engage differently than one consuming AI-generated content that was never designed with them specifically in mind. Stakeholders who have worked with you over years, who trust your judgment on what learning can and cannot accomplish, who know you will tell them the truth even when it is inconvenient — that trust is yours. AI has not built it and cannot maintain it.

Reading a Room — Virtual or Physical

The skill of sensing what is happening in a group of learners — detecting confusion that has not been voiced, discomfort that is not being named, energy that is flagging, or insight that is just forming — is fundamentally human. Skilled facilitators do this intuitively. It cannot be replicated by AI. It is the difference between a technically correct training event and one that actually moves people.

Understanding Organizational Culture and Politics

Every organization has a culture, a history, a set of unspoken rules about how decisions get made and how change is resisted. Understanding this — knowing which stakeholder needs to feel ownership before a program will get supported, knowing which topics are politically sensitive for which departments, knowing why the last three L&D initiatives failed — requires being embedded in the organization over time. This is the kind of contextual intelligence that no AI, given even detailed prompts about your organization, can fully replicate.

Making the Judgment Call When Stakes Are High

When the CEO asks you to design a leadership program for a team in crisis, and you know the real issue is not a skill gap but a trust breakdown between a leader and their team — you make a judgment call. When a compliance training deadline is about to be missed and someone suggests using unverified AI content to hit the deadline — you make a judgment call. When a client wants to automate feedback conversations that should be human — you make a judgment call. These calls require professional experience, ethical grounding, and organizational courage. AI can inform these decisions. It cannot make them.

Advocating for Learners Who Cannot Advocate for Themselves

This is the one I feel most strongly about. LXDs are, in a real sense, advocates for the learners who will sit with the content they design. Those learners are rarely in the room when decisions are made about their training. They do not get to vote on whether the compliance module is engaging or the onboarding experience is humane. LXDs who take that advocacy role seriously — who push back when content is poor, when programs are designed for organizational convenience rather than learner need, when speed is prioritized over quality — are doing something no AI can do. AI generates what it is prompted to generate. It does not advocate for the humans on the receiving end.

A Personal Roadmap for the Next 12 Months

This is the most practical section, and I want to make it genuinely specific rather than vague. Where you are in your career shapes what your next moves should look like.

Early Career LXDs (0–3 Years)

Your biggest risk is being positioned as “production support” precisely when production is being automated. Your biggest opportunity is being a native AI user who does not have bad habits to unlearn.

YOUR PRIORITIES:

• Invest immediately in learning performance consulting frameworks — Gilbert’s Behavior Engineering Model, the Ruth Clark evidence-based design principles. Develop the strategic thinking that separates junior and senior practice.
• Get serious about prompt engineering. Not just “use AI” — develop actual prompt mastery and document your best prompts.
• Pursue your first facilitation experiences. Volunteer to run workshops. Get comfortable in the room with learners.
• Find a mentor who does performance consulting, not just course development.
• Learn basic xAPI and how to think about learning measurement. Data literacy early is a massive career advantage.

Mid-Career LXDs (3–8 Years)

You have real skills but may be too production-focused. This is the moment to deliberately shift your value proposition.

YOUR PRIORITIES:

• Identify and make visible your specialty domain. What industry or skill area do you know deeply? Build content and thought leadership around it.
• Develop a “human + AI” portfolio that demonstrates strategic judgment, not just finished products.
• Seek out the stakeholder management work you may have been avoiding. Volunteer for strategic conversations, needs analyses, and business reviews.
• Build your AI governance knowledge. Read what is being published on AI ethics in learning. Develop an opinion and a framework.
• If you have not done a real ROI/impact study on a learning program, do one. Connect learning to business outcomes explicitly.
• Build or join a community of practice around AI in L&D. The eLearning Guild, ATD, and Learning & Development communities on LinkedIn are active with practitioners navigating exactly what you are navigating.

Senior LXDs and L&D Leaders (8+ Years)

Your competitive position is actually stronger than you think, but only if you actively evolve. The risk for senior practitioners is complacency — assuming that experience alone protects you.

YOUR PRIORITIES:

• Develop your AI governance and policy expertise. You are in the best position to shape how your organization uses AI in learning, but only if you do the work to understand it well enough to make informed recommendations.
• Become the person who connects L&D strategy to business strategy. Own the conversation about how AI changes what is possible in learning, but also what it cannot replace.
• Invest in facilitation mastery if you have not. Senior facilitators with genuine expertise in organizational learning are increasingly valued as async AI content proliferates.
• Build your public presence around a genuine POV on learning and AI. Write. Speak. Teach. The field needs senior practitioners who take real positions, not AI-generated summaries of neutral overviews.
• Consider developing a consulting practice or advisory capacity. The organizations that need help navigating AI in learning are numerous. The practitioners who can advise them thoughtfully are not.

Where to Learn AI Skills for L&D

RECOMMENDED RESOURCES:

• TLDC (Training, Learning, and Development Community): Active community with strong AI in L&D content.
• The Learning Guild Research: Serious research on AI adoption in L&D.
• Learnnovators Blog and ID in Focus Podcast: Thoughtful practitioner perspectives.
• Nick Shackleton-Jones’s work on affective learning design — the antidote to AI-generated content that technically covers the material but creates no emotional engagement.
• The Deliberate Practice of prompt engineering: Just use AI tools daily on real work. Nothing substitutes for actual practice.
• Coursera and LinkedIn Learning AI courses: Background knowledge on how AI works is useful even without going deep on the technical side.
• ATD’s AI in L&D resources: Growing body of practitioner-focused guidance.

The Honest Reckoning

I want to close where I opened: this is a structural shift, not a trend cycle. Some of what AI is changing is genuinely uncomfortable for practitioners who have built careers around specific execution skills. That discomfort is appropriate. It is information. It is telling you to evolve.

But here is what I also know after years in this field: the learning experience designers who have always done the most important work — the ones who diagnose before they prescribe, who advocate for learners, who build trust with organizations, who make judgment calls when stakes are high — those practitioners are not threatened by AI. They are made more visible by it, because the contrast between genuine learning expertise and AI-generated content production is getting clearer, not murkier.

The field does not need more LXDs who are fast at producing modules. It needs more LXDs who are wise about when to produce them, thoughtful about how to design them, and courageous enough to say when something else entirely is what learners actually need.

That has always been the job. AI just makes it the only version of the job that matters.

Explore more resources for evolving your practice at LXD Learning Experience Design and the LXD Career Hub.