Insights from American Educational Research Association Annual Meeting

A few weeks ago, I had the opportunity to attend the American Educational Research Association (AERA) annual meeting, a significant gathering for educational researchers and educators. At the event, several panel sessions focused on Artificial Intelligence (AI) were conducted by experts in the field of education and assessment. The discussions revolved around AI's impact on essay responses, student research, and assessment creation. While some expressed optimism about AI's potential if students understand its role, others raised concerns about biases and the black-box nature of AI systems.

Here are some key themes and takeaways from the sessions:

1) Understanding AI's Capabilities and Avoiding "Outsmarting" It

The experts cautioned against trying to "outsmart" AI by designing prompts or problems specifically to confuse it. They emphasized that AI capabilities are continuously improving, and what might seem like distinct AI characteristics now will appear more human-like in future versions. Comparisons between ChatGPT 3.5 and 4.0 were frequently cited to illustrate this point. !

2) Teaching Digital Literacy and Responsible AI Use

It was proposed that students should be taught how to use AI appropriately, particularly for research purposes. Digital literacy skills will include learning how to query AI systems effectively and assess the accuracy and relevance of their responses. Just as we teach students to create good search strings and evaluate source credibility, AI usage will become an essential component of modern digital literacy.

3) Recognizing AI's Lack of True Understanding

Despite AI's ability to produce coherent and seemingly knowledgeable responses, it was stressed that AI lacks true understanding. Personification of AI was discouraged, as its writing style might appear excellent but could miss addressing the core question. This applies to both essay writing and code generation.

4) Introducing Ethics and Implications of AI Early On

The experts strongly advocated for integrating discussions on AI's ethical implications and biases into the undergraduate curriculum from an early stage. Examples of biases found in AI systems were shared, ranging from explicit ones based on data sources to subtle ones that required closer examination. Incorporating such exercises into courses like English 101 can help students develop a critical eye towards AI-generated content.

5) Emphasizing Meaningful Engagement and Personalized Learning

The consensus was to allocate more time for in-class discussions and individualized projects that genuinely engage students. While AI has potential for personalized learning, a better understanding of its implications is crucial before fully implementing it in educational settings.

6) Acknowledging Handwritten Assignments as Not a Solution

Interestingly, no one in the sessions suggested moving towards handwritten assignments as a solution to AI challenges. Instead, the focus remained on leveraging AI responsibly and integrating it into the educational process thoughtfully.

As we move forward, it is essential to keep abreast of AI advancements, understand its limitations, and embrace its potential to enhance education while being mindful of ethical considerations. By fostering digital literacy and early discussions on AI implications, we can equip our students to navigate the AI-driven world effectively.