This year brought a shift in my thinking and interaction with AI. In previous years, I was focused on how it could replace me and marketers, and student learning. This year, I shifted more to how students could use it to learn and how marketers and I could use it to improve our jobs.
While some AI companies promise students that AI can complete homework for you, and others spend billions training AI to replace knowledge workers, I spent time focused on how AI can improve knowledge gain and enhance knowledge worker performance.
What if we trained humans to work with AI to improve learning and jobs instead of training AI to replace them? That question shaped my exploration last year. Here are 10 key insights with practical frameworks, tools, and principles you can apply now.
1 – When the Technology Leaped Forward, Multimodal AI Changed the Classroom
Early in the year, AI’s multimodal capabilities, such as voice interactions in NotebookLM and live video with Gemini 2.0, changed what’s possible in education. NotebookLM became a practical tool for creating an AI tutor trained on specific course materials. I used it in my Digital Marketing class drawing from the open source text and six trusted professional digital marketing websites.
Students asked questions and got answers with clickable citations back to source material. The Audio Overview let students interrupt AI podcast hosts to ask clarifying questions. I tested it on all assignments and in class, and it always gave accurate answers because NotebookLM draws from sources you provide, not the entire web. Students loved being able to study while listening to the AI-generated podcasts and asking questions 24/7.
Key Insight: AI tutors work best when you control the sources, and students use them to reinforce their learning, not replace it.
2 – AI Agents and Reasoning Models Arrived
The hype over AI agents and reasoning models came to a reality in 2025. Every player promised AI agents with “deep” tools such as Deep Research and Deep Search and “thinking” models. The LLMs improved with thinking mode taking more time to answer questions, but we learned not to be fooled by the names.
“Agent” implies full autonomy, which they’re not capable of—even today. AI that pauses before answering and shows its process doesn’t mean it’s thinking. It’s still a mathematical prediction machine that operates on learned patterns, not genuine comprehension. Even these advanced models don’t get it right all the time and need your context, guidance, and expertise. Yet, the truth remains AI will replace parts of your job. Upskilling is not optional. It is survival.
Key Insight: Language matters. Calling AI “thinking” or “reasoning” anthropomorphizes what’s sophisticated pattern matching. Humans need to maintain agency.
3 – Vibe Marketing: Fast Iteration Requires Deep Expertise
“Vibe marketing” sounds like winging it—the opposite of data-driven decision-making. But testing it revealed that AI can accelerate creative iteration when you have marketing fundamentals. I define Vibe Marketing as getting an idea and using AI to run with it, researching, illustrating, and iterating quickly, combining design thinking with marketing and innovation.” An in-class brainstorm exercise showed this clearly.
In one afternoon, we went from idea and white board sketch to a photo-realistic product image, logo, target market, positioning, price, place, and promotions strategy with marketing channels, ideas, and a mock Instagram ad. We even had ways to create a prototype for investors by hand, 3D printer, or rubber molding. But it only worked because of years of marketing expertise guiding every step. As Gemini confirmed: “You Definitely STILL Need Core Marketing Fundamentals.”
Key Insight: AI amplifies expertise. It doesn’t create it. The more you hand off, the more that can go wrong.
4 – AI Flattery Became a Problem
When ChatGPT-4o was updated to be more “agreeable,” it started validating everything from flat earth theory to a “poop on a stick” product idea. This is obviously dangerous. Growth comes from critique, not flattery. The most valuable feedback often hurts. Like my boss who said I “suck at presentations.” That honesty drove improvement. After a High Impact Presentations course that career weakness turned into a strength.
How do we avoid AI flattery? I created an AI Curiosity & Critique Framework. Instead of asking AI to validate ideas, use it to purposely:
- Ask divergent questions
- Challenge assumptions
- Invite dissenting viewpoints
- Validate rigorously
Key Insight: AI as a yes-person is worthless. AI as a critical thinking partner creates value.
5 – Three AI Tools That Embody Partnership
A human-first AI philosophy needs practical application. A goal for summer was to create a custom GPT. By fall, I built three examples of what human-AI partnership can look like:
This GPT gives step-by-step guidance through the audit process without doing it for you. The value of a social media audit comes from getting into each platform and observing what’s happening. It prompts strategic thinking, doesn’t replace it. In this post, I show how to create your own custom GPT.
This GPT coaches a five-act storytelling framework grounded in academic research. It helps create scripts and storyboards, but you drive story creation. Only humans have direct life experience to feel the tension and emotion that make stories authentic. In this post, I give an example of using it for a brand.
This GPT guides through segmentation, targeting, and positioning. This core marketing strategy is integral to success but often misunderstood or misapplied. It teaches the process and strengthens analytical thinking without making strategic decisions for you.
Key Insight: Each AI tool guides and augments expertise without outsourcing the cognitive work.
6 – Why Humans Remain Essential: Three Physical Brain Advantages
Beyond tactics, there’s a fundamental question: Why will humans remain essential as AI capabilities advance? The answer lies in the neuroscience of how our brains work.
- 3-D Neural Architecture vs. 2-D Grids: Human brains are messy, interconnected jungles. AI operates on 2-D grids. We can connect a novel read years ago to today’s business problem, jumping between entirely different conceptual grids. AI optimizes within its grid.
- Embodied Learning vs. Dataset Training: Crashing a minibike at age 9 teaches a thousand things about physics that no dataset can convey. This embodied learning creates intuition pure pattern matching can’t replicate.
- Energy Efficiency: Human brains run on half a banana’s worth of energy per day. Training large language models requires the power of small cities. When LLMs charge per token, those who don’t need AI for everything gain competitive advantage.
Key Insight: AI Neural networks are not structured like human brains. Use this to your advantage.
7 – For Professionals: Deepening Expertise
Based on the insights about how human brains are unique from neural networks, we can determine ways to leverage our unique brain capabilities. This only happens if you use AI to enhance thinking, not outsource it. Train your brain with the approach below.
A cognitive training approach:
- Generate alternative perspectives on strategic problems
- Reality-check domain expertise against blind spots
- Accelerate routine analysis to focus on judgment calls
- Test ideas against edge cases you might not consider
Key Insight: Maintain the “human in the loop” for synthesizing insights, making final strategic calls, understanding context and nuance, building relationships and trust.
8 – For Students: Building Brainpower
Every time AI lifts a “cognitive weight,” students miss a chance to build capabilities. Like training for a race or sports, nothing replaces hard work. Train your brain for improved learning with the approach below.
A personal trainer approach:
- Use AI as a tour guide before reading difficult texts, then do the reading
- Let AI challenge arguments, then defend and refine them
- Generate practice problems with AI, solve them without help
- Ask AI to explain concepts, then teach those concepts to someone else
Key Insight: Use AI strategically to strengthen, not weaken, your cognitive capabilities.
9- A Professional and Ethical Stance
Some AI companies are building “RL gyms” spending billions training AI to do human jobs. These training grounds are where AI learns from human experts through reinforcement learning. This creates a choice for educators and professionals: train AI to replace people, or teach people to leverage AI’s strengths to develop human capabilities.
A future worth building:
- Students develop human capabilities while leveraging AI strengths
- Professionals deepen expertise rather than outsource thinking
- Organizations value human judgment and creativity with AI speed and scale
- We treat AI as co-intelligence, not artificial human replacement
Key Insight: we can choose to help humans partner with AI instead of helping AI replace them.
10 – Design Thinking in the Age of AI
Human-centered design becomes more important as AI advances, not less. The best marketing emerges from deep human insight observed through empathy and design thinking. This final insight came from joining the Markets, Innovation & Design (MiDE) program at Bucknell University this fall.
A Human-Centered Marketing Framework:
- Insert empathy by finding the real consumer problem
- Pivot on key human insight leveraging the “aha” moment
- Make a creative leap to the big idea
- Share an engaging story for management and consumers
- Treat the marketing mix as a unified customer-centric system
This work requires human observation, human creativity, and human judgment. AI can support this approach at every stage when used correctly.
Key Insight: Deep human insight (understanding what people need and care about) remains irreplaceable.
A Path Forward
AI capabilities will continue advancing. Job displacement is real. The pressure to use AI for cost-cutting versus capability augmentation will be intense. But the outcomes aren’t predetermined. They’re being shaped right now by choices about how we use and don’t AI.
Students are discovering how to use AI as a thinking partner rather than a shortcut. Professionals are finding ways AI can free them to focus on high-value work they enjoy to deepen their expertise. As I look toward 2026, the future will continue to be built by choices to partner with AI, not be replaced by it.
How are you approaching AI in your work? Deepening expertise or outsourcing thinking?
About This Post’s Creation
This reflection demonstrates a human-AI partnership. Claude Sonnet 4.5 helped review blog posts, identify themes, and organize thoughts. But the insights, voice, and perspective come from my experience teaching, experimenting, and creating with AI while constantly asking, “What’s the best way forward?”
