← Video Library · AI Strategy
AI and the Future of Business: Opportunities, Risks, and Responsibilities
Published April 4, 2024 · 106 views on YouTube
Key takeaways
- Give AI tools like ChatGPT more context and specificity — start with the end result in mind — to get useful results instead of generic ones.
- Using AI to summarize, cite sources, and support a paper is different from generating and submitting content without editing or disclosure; the line is defined by the assignment's expectations, not a fixed rule.
- Task-focused digital work (accounting, law, copywriting) is most exposed to AI disruption; the shift is toward outcome-focused skills and strategy rather than manual task execution.
What did Darby cover in this interview about AI and business?
This is a recorded Q&A interview Darby gave to a business class at Louisiana Christian University, his alma mater, hosted by Dr. Kulp. Darby answers a series of questions sourced from the professor and students covering the good, the bad, and the ugly of AI in business — from ethical use and prompting technique to AI’s effect on the job market, academic integrity, and creative work.
How should someone start using AI in their business or day-to-day work?
Darby’s advice is to start with the simple things you already do every day and look for ways AI can save time or do them faster, rather than trying to overhaul everything at once. He describes taking an “AI first” approach: treat the AI like an assistant, give it more context and specificity about the outcome you want, and be willing to play with it and get things wrong before finding what works.
He uses the example of planning a house build: rather than asking AI to explain construction steps in the abstract, you describe the specific house you want (size, features, location) as a voice note, ask ChatGPT to act as a project manager, and iterate on the outline it produces. The same “start with the end in mind” approach applies to scheduling, market research, or any project.
Where is the line between using AI for schoolwork and cheating?
Darby says he doesn’t have a fixed answer, and that the line depends on what a given assignment is actually testing. If an assignment explicitly requires the work be done manually without AI, using it and hiding that fact is cheating. If the assignment is open about tool use (his example: “launch a startup in 24 hours, use whatever tools you want”), using AI is fair game. He recommends running AI-assisted writing through a plagiarism checker and treating AI as an assistant for research, citations, and feedback — not as a replacement for doing and editing the work yourself.
How will AI affect the labor market and specific industries?
Darby expects the biggest disruption in accounting, law, and other digitally-based, task-focused knowledge work, with jobs that rely on manual digital processes fading as AI agents take over those tasks. He contrasts this with more physical, in-person, or “uglier” businesses (he cites a friend’s mobile restroom rental trailer business as an example) that are more insulated for now because the work isn’t purely digital — though even those businesses may still lose specific functions, like copywriting, to AI tools. His broader framing: workers whose value is tied to specific tasks are most at risk, while those who can shift toward outcome-focused strategy work are better positioned.
Can an AI be customized to a specific person or business?
Yes — Darby describes this as “very customizable.” His view is that by feeding an AI your own goals, views, and how you write or make decisions, you can build a knowledge base that lets it respond and reason more like you would, and extend that same approach to building specialized agents around how you personally solve specific problems.
Full video transcript
Darby introduces the video as a shared interview with Dr. Kulp, a business professor at Louisiana Christian University — Darby’s alma mater, where he played baseball and went through the business program. Dr. Kulp invited Darby back to speak with students about the state of AI: the good, the bad, and the ugly. The interview is a series of questions sourced from Dr. Kulp and the students in the room, followed by Darby’s responses.
Dr. Kulp opens the session by introducing Darby, noting he graduated from the business department in 2015 and went on to co-found the game Side Hustle with two fellow entrepreneurs in Austin. He then turns the floor over to Darby to speak on AI’s good, bad, and ugly sides before opening it up to audience questions.
Darby explains that AI is disrupting nearly every industry, and what someone knew about AI’s capabilities six months ago may already be outdated. For businesses, this creates opportunity across HR, customer service, sales, marketing, and education — but the same speed of change opens the door to bad actors: deepfakes, unauthorized celebrity likeness use, and unresolved questions around copyright, IP ownership, and data control. He shares that he was an early customer of Jasper, one of the first companies to build on top of OpenAI’s technology, and that this experience three years prior is what led him to start Gen AI University.
He stresses thinking ethically first when integrating AI into business and personal life, since the speed at which ideas can reach the market cuts both ways — toward good outcomes or bad ones. His stated company motto is to “adapt, integrate, and grow”: using AI to free up time from routine tasks (like writing a blog post) and redirecting that capacity toward higher-leverage work. He recommends people practice consistently, even without a specific goal in mind, framing it as a way to stay ahead of a widening gap between those who adopt AI tools and those who don’t.
Dr. Kulp asks Darby to describe what “talking back and forth” with AI looks like in practice. Darby compares a large language model to an assistant across the table: vague or poorly specified requests get poor results, while providing more context — what he calls “priming” the AI — gets more useful, targeted output. He illustrates this with an ice cream analogy: asking for “ice cream” gets 500 generic options, while specifying the flavor and origin narrows it down to what you actually want. He frames this as a skill that improves with practice, since these tools are designed to be conversational.
Asked about Gen AI University, Darby describes it as an AI education community and platform for creators, founders, entrepreneurs, business owners, and marketers, offering expert interviews, courses, membership, a newsletter, and a YouTube channel, running for about three years at the time of the interview.
Dr. Kulp asks how AI could help with a project like building a house. Darby again applies “start with the end in mind”: describing the house you want (size, bedrooms, pool, location, amenities) as a voice note to ChatGPT, framed as a project manager for home construction, then iterating on the step-by-step breakdown it generates — including finding contractors — before adjusting the plan further. He notes the same approach applies to any project, from buying a car to planning content.
Asked how ChatGPT relates to what he teaches at Gen AI University, Darby explains that ChatGPT is one interface from OpenAI, one of the leading large language model companies, and that his business helps people use it (and other tools, including Claude from Anthropic) strategically rather than teaching any single tool in isolation. He mentions Gen AI University has built some of its own AI tools on top of the OpenAI and Anthropic APIs.
A student asks how to use AI for writing papers without it becoming cheating. Darby says he doesn’t have a complete answer, but that it depends on what a given assignment is testing — if AI use isn’t allowed and a student disguises using it, that’s dishonest; if the assignment is open to any tools, using AI is fair. He notes that teachers who succeed will be the ones who define clear expectations around AI use and teach students to use it ethically, rather than simply banning it or letting students disengage entirely.
A follow-up question asks whether AI would put a quotation like the opening of the Gettysburg Address in quotation marks. Darby says it can if instructed to, but this is where plagiarism-checking tools and manual review become important — AI-assisted writing should still be reviewed, edited, and properly cited, not submitted as-is.
Asked about the best way for a student to study AI to enter the AI job market, Darby first clarifies whether the question is about becoming an AI engineer or about using AI within a job. Given the engineering framing, he points to free courses from major universities, learning tools like Make.com for no-code/low-code automation, and reading API documentation from providers like OpenAI, alongside YouTube tutorials, as ways to get started.
A student asks whether AI can be given a person’s goals and manage their schedule. Darby says this is a growing use case (AI agents) — compiling your schedule, routines, and goals into one document (e.g., a Google Doc) and giving that context to ChatGPT so it can help build and adjust a schedule. He extends this into a broader vision of “swarms” of AI agents that pass tasks to each other, describing his own business’s market research workflow, where research that once took three to five hours of manual work now takes about 30 minutes using a system of connected AI assistants — a workflow he says has taken him roughly a thousand hours to build over three years and now underpins a paid research report product for his business.
Asked how AI will affect the labor market, Darby names accounting and law as fields likely to change significantly, along with other digitally-based, task-focused knowledge work. He contrasts this with more physical or in-person businesses (using a mobile restroom rental trailer business as an example) that are currently more insulated, though even those may lose specific functions like copywriting to AI tools. His framing is that workers should shift focus from task execution to outcomes and strategy.
Asked how customizable AI can be to an individual, Darby says it can be built around a person’s specific goals, views, and decision-making style by feeding it that context as a knowledge base, extending to building specialized agents that reflect how a specific person approaches problems.
Asked whether AI will replace programmers, and about AI in art, Darby says programming is one of the more replaceable roles, though understanding what good code looks like still matters for reviewing AI-generated code. On art, he shares an anecdote about a student’s AI-generated fishing image that his daughter immediately identified as AI-made, and predicts today’s AI-generated images will be nearly indistinguishable from human-made work within a couple of years. He believes people who already have artistic skill and understanding will get better results from AI art tools than those without that background, and raises open questions around how artists will be compensated as AI art scales.
On uses of AI in education, Darby points to curriculum development as the most common teacher-side use he has seen, along with reviewing and grading assignments, and highlights brainstorming and articulating thoughts as key AI use cases for students.
Dr. Kulp closes the session by thanking Darby, noting the class discussion has given him ideas for his own research course, and Darby offers to share digital and physical copies of his upcoming book on market research once it’s ready.