AI-powered carnivores: An active learning template

Post: HA-AIA; Example AIG-HE; Generative AI used: Anthrop/c Claude2, on July 20, 2023. This content is made available under the Creative Commons CC-BY-NC-SA 4.0 License (Attribution-NonCommercial-ShareAlike, View License Deed | View Legal Code).

The disruptive presence of AI in higher ed means that recently I have been thinking a lot about new ways to engage students with this new technology, and adapt our teaching approaches to not only build subject-matter knowledge and skills, but also to help students develop future-ready, AI-related skills. 

In this post I want to share an idea – and an example – for an active learning activity that uses AI to build the knowledge of a specific subject-matter topic, but also has a secondary goal of helping students develop AI skills in a way that is openly acknowledged, ethical, and where using AI judiciously is built into the assignment.

As an instructor I would start by using ChatGPT and online search engines (Google or AI-powered Bing) to identify a few scholarly articles suited to my chosen topic (in the example below, the topic is “State of Research on Carnivore Diets”. I’d ask ChatGPT to generate short summaries of the “candidate articles” to help select the most relevant sources and narrow down the list of “candidates”.

For each selected article, I would use Claude to produce various article-related resources, starting with an APA/MLA-formatted reference list, then an outline, a 200-word summary, and finally a 10-question multiple choice quiz with annotated answer key, linked back to supporting text in the text of the article. This approach structures each article as a self-guided mini lesson blueprint. 

The relatively new building in which I, my team, and over 80 faculty work, has several active learning classrooms, each seating 180 students at 20 round tables of 9. I would split each table into groups of 3 (or two, as needed), assigning each group a different article (found, selected and prepared for this activity as outlined in the paragraph above), with a brief outline and summary, and a 10-questions quiz (without the answer key and text references provided) to review and discuss. Using Claude’s estimates, which are based on the average estimates of human reading and information processing speed, I would allot appropriate time for reading the article and completing the quiz, with each group of 3 working as a team.

After this individual (reading) and group work (quiz, discussion of answers, ad identification of sections in the article that support each answer selected), the 3 groups at each table would present their article’s key findings / takeaways to the other two groups. Then, the entire table would collaboratively produce a 100-150 word executive brief summarizing the state of research on this particular topic. 

As a take-home analysis assignment, I would provide students with an “executive brief” independently generated by Claude or ChatGPT on the same topic. Working as a follow-up to in-class activity, students would be asked to compare the two summaries by analyzing similarities, differences, and details covered or missing in a brief online discussion post. 

As follow-up, second phase of this activity, I would have students at each active learning classroom table re-create and re-trace this process for a new topic, using the AI tools, with clear instructions, resources, and deadlines. This way they would learn how to gather and evaluate information, and how to use both their own cognition and available AI tools to get up to speed on any topic in their professional and personal lives. 

Guiding students to use emerging technologies effectively is important, if not critical. This activity blueprint integrates simple collaborative learning strategies and AI skill-building to produce engaged, adaptable graduates ready to thrive in our AI-driven future.

Below, I have included a sample article file I would generate for class focused on “Research on Carnivore Diets.” All content below, generated by Claude2 (AIG-HE)

APA and MLA format citations for the article: 

• APA Format: Lennerz, B.S., Mey, J.T., Henn, O.H. et al. (2021). Behavioral characteristics and self-reported health status among 2029 adults consuming a “Carnivore Diet”. Current Developments in Nutrition, 5(12), nzab133. https://doi.org/10.1093/cdn/nzab133
• MLA Format: Lennerz, Belinda S., et al. “Behavioral Characteristics and Self-Reported Health Status among 2029 Adults Consuming a “Carnivore Diet”.” Current Developments in Nutrition, vol. 5, no. 12, 2021, p. nzab133., https://doi.org/10.1093/cdn/nzab133. Accessed date accessed. 

Classroom timing estimates:

• The article contains approximately 3,500 words. The average college student reads around 200-400 words per minute. To read carefully and fully comprehend the material, let’s estimate on the lower end of that range, 250 words per minute. At a rate of 250 words per minute, it would take the average college student around 14 minutes to carefully read through the 3,500 word article. 

So, in summary: – Word count of article: ~3,500; Average reading speed: 250 words/minute; Estimated time to read carefully: 14 minutes

This is just an estimate based on average reading speeds. The actual time could vary depending on the individual’s reading ability and how carefully they read and comprehend the material. But an average college student should expect around 14 minutes to read through the full article at a pace that allows for good understanding.

Quiz time estimate

Here is an estimate for how long it would take the average college student to complete the 10 question multiple choice quiz I generated based on this article:

• Each multiple choice question with 4 answer options should take around 1 minute to read and select an answer. There are 10 questions in the quiz. So just for answering the questions, it should take about 10 minutes. However, the questions reference details from the article, so the student would likely need to go back and check the article for relevant information to confirm the right answers. Having to look back through the 3,500 word article could double or triple the base 10 minutes needed just for answering. To account for checking the article to confirm answers, let’s estimate 20-30 minutes total to complete the 10 question quiz.

In summary:

• Estimated time per question: 1 minute; Number of questions: 10; Base time for answers alone: 10 minutes; Total estimated time including checking article: 20-30 minutes. So I’d expect the average college student to spend around 20-30 minutes in total to carefully complete the 10 question multiple choice quiz, including time to reference the original article to confirm the accuracy of their selected answers. 

Summary of the article

Bullet Point Outline:

• Study surveyed over 2000 adults following a carnivore diet for at least 6 months. Diet was mostly red meat, eggs, and dairy, with little to no plant foods. 
• 93% chose the diet for health reasons, especially weight loss, improved mental health, and resolving digestive issues.
• Reported benefits included weight loss, improved mood, more energy, better sleep, and resolution of chronic health conditions like diabetes and autoimmune disorders.
• Adverse effects were very low overall. Some GI issues, muscle cramps, hair loss in <5% of people. 
• Bloodwork showed increased LDL but optimal HDL and triglycerides. Marked A1c and medication reductions for diabetics.
• High diet satisfaction and no major social impact reported.

200 Word Summary: 

This study surveyed over 2000 adults who had followed a carnivore diet for a minimum of 6 months. The diet consisted predominantly of red meat, eggs, and dairy, with less than 10% consuming plant foods like vegetables, fruits or grains more than monthly. 93% of respondents chose the diet for health reasons, especially weight loss, mental health, and digestive issues. Participants reported benefits including weight loss, improved mood, energy, sleep and resolution of chronic health conditions like diabetes and autoimmune disorders. Adverse effects were uncommon, with some gastrointestinal issues, muscle cramps and hair loss occurring in less than 5% of respondents. Bloodwork changes included increased LDL cholesterol but optimal HDL and triglycerides. For diabetics on the diet, marked reductions were seen in HbA1c and diabetes medications. Overall, participants reported high diet satisfaction and little social impact from following the carnivore diet. The study is limited by self-reported data and likely biases in this selected cohort. But it provides initial insights into perceived health effects of a meat-based, plant-free diet pattern. Long term randomized trials are needed to further assess the diet’s safety and efficacy.

Here is a 10 question quiz in the requested format:

Question 1: What percentage of surveyed participants reported choosing the carnivore diet for health reasons?

A) 33% B) 45% C) 73% D) 93%

Answer: D Quote: “93% of participants stated health reasons as their motivation for beginning the diet.” (page 3)

Question 2: Which of the following were commonly reported benefits of the carnivore diet?

A) Improved mood B) Weight loss C) Better sleep D) All of the above 

Answer: D (Blog author note: Claude2 tends to include “All of the above” on the quizzes it generates. Overall, it performs much better on quiz generation than ChatGPT-4 Plus, but if I were to use this quiz in a real classroom, I’d change this to “All answer choices proposed” (or similar), and I would shuffle the order of potential answers/distractors. As I noted elsewhere, I think at this stage of generative AI ability, such quizzes are useful as practice or review quizzes, or as a basis for an in-class activity, but are not yet ready to be used as actual assessment quizzes, unless they have been very thoroughly reviewed and edited.)

Question 3: What percentage of participants reported new or worsened diarrhea after starting the carnivore diet? 

A) 1.1% B) 3.1% C) 5.5% D) 7.2% 

Answer: C Quote: “New or worsened diarrhea occurred in 5.5%.” (Supplemental Table 1)

Question 4: How did the carnivore diet affect HbA1c and diabetes medication use in diabetic participants?

A) No significant changes were seen B) Modest reductions were reported C) Marked reductions were reported D) Increases were reported

Answer: C Quote: “Respondents with diabetes experienced relatively large median IQR reductions in BMI (4.3 1.4–7.2) and HbA1c (−0.4% 0% to 1.7%)…Diabetes medication use was significantly reduced.” (page 7)

Question 5: What percentage of participants reported consuming vegetables more than once per month? A) Less than 10% B) 20% C) 37% D) Over 50%

Answer: A Quote: “Less than 10% of respondents consumed starchy vegetables, nonstarchy vegetables, fruits, or grains more often than once monthly.” (page 3)

Question 6: What percentage of participants reported taking vitamin supplements?

A) 10% B) 37% C) 63% D) 75% 

Answer: B Quote: “37% reported no use of any vitamin supplements.” (page 3)

Question 7: What was the most commonly consumed food in the carnivore diet?

A) Eggs B) Pork C) Poultry D) Red meat Answer: D

Question 8: What blood lipid changes were observed in participants on the carnivore diet? A) Increased HDL cholesterol B) Decreased triglycerides C) Increased LDL cholesterol D) All of the above

Answer: D 

Question 9: What percentage of participants with type 2 diabetes discontinued insulin after starting the carnivore diet? 

A) 84% B) 90% C) 92% D) 98% 

Answer: C Quote: “92% of participants with T2DM discontinued insulin” (Table 3)

Question 10: How did participants rate their overall health after following the carnivore diet?

A) Mostly worse B) Unchanged C) Mostly better D) Much worse

Answer: C Quote: “Participants reported improvements in chronic medical conditions, general health, and aspects of well-being such as energy, sleep, strength, endurance, mental clarity, memory, and focus.” (Figure 1)