Higher Ed AV Media is proud to introduce AI Insights with Andy a fresh, thought-provoking column by Andy Vogel that explores how artificial intelligence intersects with teaching, learning, and the daily realities of higher ed technology. Drawing from hands-on experiences, Andy blends relatable stories with practical takeaways. Each installment offers an authentic, grounded look at where AI helps, where it falls short, and how human judgment and creativity remain essential.
Andy Vogel is an Instructional Design Specialist at The Ohio State University where he blends emerging tech with hands-on learning. His recent work includes VR collaborations projects that explore how immersive tools can enhance engineering education. Andy’s focus is on making complex concepts accessible—and fun—through creative design and strategic tech integration
Connect with Andy on LinkedIn
Check Out Andy’s Digital Portfolio
Why Validating AI Matters
By Andy Vogel
In an era where artificial intelligence is increasingly embedded in our workflows—from writing assistance to student advising—it’s easy to fall into the trap of assuming that what AI generates is always correct, complete, or even appropriate. AI’s rapid emergence of convincing information can alter our sense of media literacy. These factors can create both environmental and digital pollution.
As with musical performance, AI outputs can be technically accurate yet emotionally off-key. That’s why validation—the act of critically reviewing and refining AI-generated content—is not just a technical necessity, but a human responsibility.
As someone who works at the intersection of education and technology, I’ve found an unexpected ally in this process: music. Specifically, my experience playing bass and guitar helps me think about the tone, rhythm, and harmony of AI outputs. While I am no Jimi Hendrix, I can tell when the notes or tablature generated by AI are way off.
Music is Valid
To take a break from doomscrolling, I wanted to play doom metal. I started testing a bass at a local shop and knew that was my plan. Before I made the purchase, I consulted AI on the best deals for a full setup. Here’s the output when I asked it to compare two local models I found:
⚡ Dean Metalman 4-String Bass
Pros:
- Metal-Focused Design: Specifically designed for metal genres, featuring a high-output humbucker near the bridge for a punchy tone.
- Affordability: Generally more budget-friendly, making it accessible for beginners.
- Unique Aesthetics: The ML or V body shapes offer a distinctive, aggressive look. (Sources: Chicago Music Exchange, zZounds, Reddit)
Cons:
Ergonomics: Some users report neck dive and discomfort when playing seated due to the body shape. (Sources: Basschat, Reddit)
Limited Versatility: The single-pickup configuration may not offer as much tonal variety.
🏆 Recommendation
If you prioritize build quality, tonal versatility, and a solid foundation for various genres, the Washburn B-2 MIJ is a superior choice. Its craftsmanship and upgraded bridge make it a reliable instrument that can grow with you as a player.
However, if you’re seeking an affordable, metal-centric bass with a striking appearance, the Dean Metalman fits the bill. It’s tailored for heavy genres and offers a straightforward setup ideal for beginners.
Final Tip: If possible, try both instruments to see which feels more comfortable and suits your playing style. Personal preference plays a significant role in choosing the right bass.
The LLM pulled from Reddit to outline its recommendations. I went in person to the store and played each to make my own judgment. I used the same process to choose the amp.
🎸 Comparison: Behringer BXL450 vs. Peavey Basic 60
| Feature | Behringer BXL450 | Peavey Basic 60 |
| Power Output | 45 Watts | 60 Watts |
| Speaker Size | 10″ Bugera | 12″ |
| Channels | Clean and Distortion | Typically Clean Only |
| EQ | 5-band graphic EQ with FBQ Spectrum Analyzer | 3-band EQ |
| Headphone Output | Yes | Yes (via line out) |
| Additional Features | CD input, opto-limiter, wedge-shaped cabinet | Basic features |
| Portability | More compact and lighter | Slightly larger and heavier |
I selected the Washburn with the Behringer amp. Then it was time to start learning songs.
Metallica was in town and played a rendition of the Buckeye State favorite, Hang on Sloopy. I knew there was no official tab because this was an off-the-cuff performance, but AI was up to the task. From my chats, you can see that I had to prod the AI to find the tab. What likely happened is that it found the original version by The McCoys and reworked it into a heavier style.
The chat also offered to simplify the tab and recommended “light slaps” on the intro chords.
Comparing AI Outputs to Musical Interpretation
When I play music, I’m constantly making micro-adjustments: am I too late or early? Does this note need more attack? Is my tempo drifting? I ask these questions not only to decide if I can play the song but also to check if the tab provided by AI is correct. I already knew the rhythm of Hang on Sloopy, so I could play by ear and cross-check against the tab.
We know AI often scrapes the web for quantitative social media engagement rather than qualitative peer reviews. So it’s not surprising that ChatGPT pulled tabs from forums—many of which are crowd-sourced. Premier websites offer paid, official versions of songs in both sheet and tab form. As a layman, I stuck with the AI-generated tabs and used my playing to determine their accuracy.
Bias, Nuance, and Human Judgment
I employed the Plus version of OpenAI’s ChatGPT as my “instructor.” The Plus version included access to other GPTs such as Guitar Tablature, but I did not find it accurate or capable of finding certain songs.
Through this process, I’ve learned to:
- Spot biases AI might unintentionally reinforce—such as assumptions about musical style.
- Recognize when nuance is missing, especially in unlisted covers and remakes.
- Trust my ear when something “sounds wrong,” even if I can’t immediately explain why.
Just as a musician interprets a score differently than a machine might, we must interpret AI outputs with a critical, creative ear.