According to Gram Research analysis, ChatGPT can accurately measure how easy nutrition articles are to read and whether they sound positive or negative, performing just as well as human readers. A 2026 study published in JAMIA Open found no significant differences between ChatGPT and human coders when analyzing reading level and sentiment in magazine nutrition articles, suggesting AI tools could help editors make health information more accessible to different audiences.
A new study tested whether ChatGPT, an artificial intelligence tool, could accurately measure how easy nutrition articles are to read and whether they sound positive or negative. Researchers analyzed nutrition articles from magazines over one year and compared ChatGPT’s analysis to human readers. The results showed that ChatGPT worked just as well as humans at measuring reading difficulty and sentiment. Most nutrition articles were written at fairly high reading levels, and articles written in more complex language tended to sound more positive. This research suggests AI could help editors make nutrition information more accessible to different audiences.
Key Statistics
A 2026 study in JAMIA Open found that ChatGPT performed just as well as human readers when measuring reading level and sentiment in nutrition articles, with no statistically significant differences between AI and human analysis.
In a 2026 analysis of nutrition articles from legacy media, the most common topics were food and diet recommendations for health (23.8% of articles), celebrity diet stories (15.6%), and cooking articles (14.1%).
A 2026 study found that nutrition articles written at higher reading levels showed more positive sentiment values than articles written in simpler language, suggesting complexity and tone are related in nutrition journalism.
The Quick Take
- What they studied: Whether ChatGPT could accurately measure how easy nutrition articles are to read and whether they sound positive or negative, compared to human analysis
- Who participated: Nutrition articles from magazines collected over one year, analyzed by both ChatGPT and human coders
- Key finding: ChatGPT performed just as well as human readers at measuring reading difficulty and sentiment in nutrition articles, with no significant differences between AI and human analysis
- What it means for you: Magazine editors and health communicators could use AI tools to check whether their nutrition articles are too complicated or hard to understand before publishing them, making health information more accessible to everyone
The Research Details
Researchers collected nutrition articles from magazines over a one-year period and analyzed them in two ways. First, they used ChatGPT, an AI tool, to measure the sentiment (whether articles sounded positive, negative, or neutral) and reading level of each article. Second, they used the SMOG formula, a standard method for measuring how difficult text is to read. They also had human readers analyze some of the same articles to see if ChatGPT’s results matched human judgment.
The researchers then organized the articles by topic to see if different types of nutrition content had different reading levels or sentiments. They looked at articles about celebrity diets, cooking tips, food recommendations for health, and food insecurity, among other topics. This approach let them understand both how AI performs compared to humans and whether different types of nutrition content are written differently.
This research matters because nutrition information is important for public health, but if articles are too complicated or written in confusing ways, people won’t understand or use the information. If AI tools like ChatGPT can accurately measure reading difficulty and tone, editors could use these tools to improve how they communicate about nutrition. This could help make health information more accessible to people with different reading abilities and backgrounds.
The study’s strength is that it directly compared ChatGPT’s analysis to human coders and found no significant differences, suggesting the AI tool is reliable for this task. However, the study did not specify the exact number of articles analyzed or provide detailed information about the human coders’ qualifications. The research was published in JAMIA Open, a peer-reviewed journal focused on health information technology, which adds credibility to the findings.
What the Results Show
ChatGPT performed just as well as human readers when measuring both reading level and sentiment in nutrition articles. This means the AI tool could be trusted to do this type of analysis without human involvement. The study found that nutrition articles in magazines had relatively high reading levels overall, meaning they were written in fairly complex language. Articles written at higher reading levels tended to have more positive sentiment, or tone, compared to articles written in simpler language.
The researchers identified twelve different types of nutrition content in the articles. The most common topics were food and diet recommendations for health (appearing in 23.8% of articles), celebrity diet stories (15.6%), cooking and culinary articles (14.1%), and articles about food insecurity (10.7%). Different types of articles had different reading levels and sentiments, which makes sense because magazines target different audiences.
The study found that sentiment ratings across all articles stayed close to neutral on average, with a mean score of 0.26 on a scale where neutral is zero. This suggests that nutrition articles in magazines don’t typically use strongly positive or negative language. The reading levels and sentiment varied depending on which magazine published the article and what topic the article covered, which indicates that different publications write for different audiences and different nutrition topics are discussed in different ways.
This is one of the first studies to test whether AI tools like ChatGPT can analyze nutrition articles specifically. Previous research has shown that AI can measure reading level and sentiment in other types of health content, but this study is novel in applying these tools to nutrition journalism. The findings support the broader idea that AI tools are becoming reliable for analyzing written health communication.
The study did not report the exact number of articles analyzed, which makes it hard to understand how large the research was. The study also did not provide details about which magazines were included or how articles were selected, so it’s unclear whether the results apply to all nutrition coverage or just certain publications. Additionally, while the study showed ChatGPT works as well as humans for this task, it didn’t test whether using these AI tools actually leads to better nutrition articles or whether readers find them more helpful.
The Bottom Line
Magazine editors and health communicators should consider using ChatGPT or similar AI tools to check the reading level and tone of nutrition articles before publication. This could help ensure articles are appropriate for their target audience. However, AI analysis should be used as a helper tool alongside human judgment, not as a replacement for it. Editors should aim to match reading levels to their audience and consider whether the tone of articles matches the topic and audience needs.
Magazine editors, health journalists, nutrition communicators, and anyone responsible for writing or publishing nutrition information should care about this research. People who struggle to understand health information because articles are too complicated should also care, as this research could lead to more accessible nutrition content. Healthcare providers who recommend articles to patients might also benefit from knowing which articles are written at appropriate reading levels.
If editors start using AI tools to analyze and improve their nutrition articles, readers could see more accessible content within a few months to a year. However, changing how an entire industry writes takes time, so widespread improvement in nutrition article readability might take several years.
Frequently Asked Questions
Can ChatGPT measure how hard nutrition articles are to read?
Yes, according to a 2026 study in JAMIA Open, ChatGPT can accurately measure reading difficulty in nutrition articles just as well as human readers. The AI tool performed with no significant differences compared to human analysis.
Are nutrition articles in magazines too complicated for most people?
A 2026 study found that nutrition articles in magazines have relatively high reading levels overall, meaning they’re written in fairly complex language. This varies by magazine and topic, but suggests some readers may find them difficult to understand.
What types of nutrition articles appear most in magazines?
A 2026 analysis found that food and diet recommendations for health were most common (23.8%), followed by celebrity diet stories (15.6%), cooking articles (14.1%), and food insecurity articles (10.7%).
Should magazines use AI to check their articles before publishing?
A 2026 study suggests AI tools like ChatGPT could help editors measure reading level and tone in nutrition articles. Using AI as a helper tool alongside human judgment could make nutrition information more accessible to different audiences.
Want to Apply This Research?
- Track the reading level of nutrition articles you read each week using the SMOG score. Note whether you found articles easy or difficult to understand, and compare this to the reading level measurement. Over time, you’ll learn which reading levels work best for you.
- When searching for nutrition information, use the app to filter articles by reading level. Start with articles rated at lower reading levels if you want quick, easy-to-understand information, or choose higher reading levels if you want more detailed, technical information.
- Keep a log of nutrition articles you read and rate how easy they were to understand on a scale of 1 to 10. Compare your ratings to the AI-measured reading level to see if the measurements match your experience. This helps you learn which reading levels work best for you and helps improve the AI tool’s usefulness.
This research demonstrates that AI tools can measure reading level and sentiment in nutrition articles, but it does not provide medical or nutritional advice. The findings about specific nutrition topics (such as celebrity diets or food recommendations) should not be interpreted as endorsements or criticisms of those approaches. Always consult with a qualified healthcare provider or registered dietitian before making changes to your diet or nutrition habits based on any article, regardless of its reading level or tone. This study shows that AI can be a useful tool for editors, but human expertise in nutrition and health communication remains essential.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.