Whether it’s for a homework assignment, job interview, or studying, we can find any answer using AI. But the answer is not synonymous with knowledge.
Traditionally, knowledge is rooted in human cognition: It’s something people possess and develop through learning and reflection. We further our knowledge by seeking experiences and ideas that challenge our thinking.
Yet the rise of artificial intelligence has called this human-centric view into question. AI, especially natural language models like ChatGPT, is beginning to challenge and expand our definitions of knowledge — and not necessarily in a good way.
ChatGPT, launched by OpenAI in 2022, has skyrocketed in popularity, especially with students. Today, the program has over 180 million users. AI models like ChatGPT can process, generate, and summarize from multiple sources instantly, offering users information on everything from history and science to specialized fields like medicine, law, and engineering.
However, people — especially students — must understand that there is a difference between reading answers and truly comprehending information. The ability to distinguish between information and knowledge becomes increasingly important as AI systems continue to make mistakes and reflect biases.
These AI systems have limitations and inaccuracies, producing flawed outputs. Second-year student Kara Kueper, a chemistry major at the University of Iowa, used ChatGPT to study for tests, only to discover its errors.
“ChatGPT doesn’t do too well for science. I tried having it explain and solve practice physics problems, and almost all the answers were wrong,” she explained.
“But trying to get the AI to do work for any writing assignment is almost worse,” Kueper added.
Writing is innately a human concept. AI provides a wealth of information, but humans must apply context, reasoning, and critical thinking to turn that information into reliable and actionable knowledge or writing. Relying on ChatGPT comes with many ethical costs, starting with changing the definition of knowledge.
Recent studies and observations have shown AI can perpetuate discrimination, especially against women. Trained on large datasets, often collected from biased real-world sources, AI systems learn from and replicate this bias. For example, Amazon’s chatbots were instructed to review job applications, but instead of taking a gender-neutral approach, the system was influenced by the male-dominance in the tech field.
ChatGPT works in a similar way. Depending on the prompt and context, it has been found to produce gender biases and stereotypes. UI first-year student Emily Carmona noticed this pattern when seeking help from ChatGPT.
“I’ve had it generate text where the gender stereotypes are blatant — like automatically assigning roles like ‘nurse’ to women and ‘doctor’ to men,” she explained.
This lack of objectivity may stem from the creators themselves.
The reinstallation of OpenAI’s CEO Sam Altman brought many changes, including the emergence of a new board. Unlike the previous members, this board is composed entirely of men. Research suggests that the exclusion of women reduces attention to ethical and safety concerns regarding artificial intelligence. When diverse perspectives are absent, AI technologies risk reflecting narrow viewpoints, leading to systemic inequalities.
Consequently, knowledge may no longer be something developed through new information and experiences. Rather, it is increasingly associated with the retrieval of facts — biased or not.
Disconnection from intellectual and emotional processes will perpetuate cycles of prejudice.