Artificial Intelligence:
Critical Evaluation

Generative AI is powerful and can accelerate research, drafting, and revision.  However, its convenience introduces a range of ethical challenges that every researcher should consider.

• Academic integrity
  Higher education institutions increasingly express concern about how AI blurs the line between original work and plagiarism. When AI tools produce text that is incorporated into student work, it undermines the principle of original scholarship and weakens the development of academic skills. 

• Data privacy
  One of AI's most noteworthy strengths is its ability to process a large amount of data quickly, but this strength raises privacy concerns. When sensitive or personally identifiable information is collected and used for training, individuals may lose control over their own information. Data stewardship is essential. 

• Influence on student learning
  AI's shortcuts can impede the development of critical thinking. Cummings (2024) warns that "some future AI approaches... could constrict the questions researchers ask, the experiments they perform, and the perspectives that come to bear on scientific data and theories. " Overreliance on AI risks weakening intellectual curiosity. 

• Hallucinations and unreliability
  Though AI responses sound confident, AI systems are known to hallucinate, producing inaccurate information, misinterpreting data, and fabricating citations. These errors have serious consequences, requiring a high-level of oversight.  

• Bias from trained data
  AI's trained data is inherently biased though subsequently hard to detect. This bias systematically skews responses to prompts and leads to prejudicial perceptions based on the biased outcomes. 

When working with GenAi, oversight means actively verifying the accuracy of AI-generated content. One of the most reliable and simple methods is through lateral reading.  Lateral reading is the practice of checking multiple independent sources to confirm whether information is credible. This technique is useful for spotting misinformation anywhere, but it's especially valuable when evaluating AI output.