The Rich Get Richer in the AI Game
Willy A Renandya, 30 March 2025
Many people think artificial intelligence is a great equalizer, a tool that gives everyone the same advantages. But the truth is more complicated. Just as wealth tends to grow faster for those who already have money, knowledge expands more quickly for those who already understand a subject well.
When it comes to using AI, this means the gap between experts and beginners may actually grow wider instead of shrinking. Thus, the adage “the rich get richer, the poor get poorer” applies here.
This effect shows up clearly in education, where AI tools are increasingly used to create lesson plans, tests, and study materials. Teachers with deep knowledge of their subjects can use AI to produce excellent resources quickly. Those with less knowledge and experience often get poorer results, not because the technology fails them, but because they lack the expertise to use it effectively.
Knowledge matters in giving clear prompts
AI systems respond best to precise, well-structured requests. People who know a subject thoroughly can give the AI detailed instructions that lead to useful outputs. Those without the same background often ask vague questions and receive vague answers in return.
For example, consider two teachers requesting AI assistance with a geography lesson. A new teacher might simply ask for “a lesson about Southeast Asian climate zones.” The AI would likely produce a superficial overview listing basic climate types like tropical monsoon and equatorial, with little practical application for the classroom.
An experienced geography teacher, on the other hand, might specify: “Create a 90-minute lesson for 10th graders analyzing how urbanization in Jakarta, Bangkok and Manila affects their microclimates. Include comparative satellite images from 1990-2020, data interpretation exercises on urban heat island effects, a group discussion about sustainable city planning solutions, and a reflection connecting these issues to students’ own communities.” The AI will generate a comprehensive, ready-to-adapt lesson plan complete with visual aids, activities and local relevance.
In language assessment, the difference is equally striking. A new language teacher might simply ask AI to “create a speaking test,” resulting in generic questions like “Describe your hometown” with no clear scoring system.
An expert in language teaching would provide specific instructions: “Design a paired speaking task for intermediate English learners to assess interactive communication skills. Include a problem-solving scenario like planning an event together, clear instructions for test-takers, a 4-point scoring rubric focusing on fluency, interaction strategies, grammar accuracy and vocabulary range, plus sample responses for each score level.” The AI generates a complete, ready-to-use assessment that matches professional standards.
Knowledge helps in spotting and fixing mistakes
AI sometimes provides information that sounds correct but isn’t, or leaves out important details. People with expertise in a subject can identify these problems immediately. Those without the same knowledge might not notice anything wrong.
Consider a science teacher using AI to create a quiz about ecosystems. A new teacher might accept all the questions the AI generates, even if some are poorly worded or cover topics the class hasn’t studied. An experienced teacher would review each question, remove ones that don’t fit, add better ones, and adjust the difficulty level to match what students have learned.
This ability to evaluate and improve AI’s work makes all the difference in the quality of the final product. Without it, people may unknowingly use materials that are incomplete, inaccurate, or inappropriate for their needs.
From first drafts to finished products
AI provides a solid first draft, but rarely a finished product. Experts use their knowledge to edit, expand, and tailor what the AI produces. Beginners often use the first version without changes, missing opportunities to make it better.
For instance, an AI might generate a decent set of slides about Shakespeare’s plays. A literature teacher with years of experience would rearrange the slides to tell a clearer story, add memorable examples from different productions, and include discussion prompts that get students thinking critically. A less experienced teacher might present the slides exactly as the AI created them, resulting in a lesson that works but doesn’t shine.
How to close the gap
When only some teachers can use AI effectively, it creates uneven opportunities for students. Schools with many expert teachers will have constantly improving, high-quality materials. Those with mostly new teachers may rely on weaker AI outputs that never get properly refined.
There’s also a risk that beginning teachers will trust AI too much instead of developing their own teaching skills. If new teachers always use AI materials without learning how to create their own, they may never gain the expertise needed to use the technology well.
Making AI truly helpful for everyone requires several steps. First, training should teach people not just how to use AI tools, but how to evaluate and improve what those tools produce. Second, schools and other organizations should encourage collaboration between experts and beginners, so knowledge gets shared. Finally, there should be systems to check the quality of AI-assisted work, making sure it meets proper standards before being used.
Conclusion
AI won’t automatically make education or other fields more equal. Like any powerful tool, it works best for those who already know how to use it well. Without conscious effort, AI could make existing knowledge gaps grow even wider. The challenge isn’t just adopting the technology, but making sure everyone can benefit from it equally.
The solution lies in recognizing that AI is a partner, not a replacement, for human expertise. The more we value and develop real understanding in people, the better our AI tools will serve everyone.
More readings
2 Replies to “The Rich Get Richer in the AI Game”