Develop Research Topics
Generative AI tools, like Chat GPT, can help you brainstorm and narrow down your research topics. For instance, if you are interested in a general topic of mental health, Generative AI can break it down into smaller aspects that may help you determine a more specific topic.
Example: When asked for a more specific research topic about mental health, Chat GPT provided the following suggestions:
Mental Health Stigma: Study societal attitudes impacting help-seeking behaviors and ways to reduce stigma.
Therapeutic Approaches: Compare the effectiveness of therapeutic interventions like CBT, DBT, and mindfulness-based therapies.
Specific Populations: Examine mental health in specific groups (children, adolescents, elderly, LGBTQ+, refugees).
Technology's Impact: Investigate how technology, social media, and screen time affect mental health.
Workplace Mental Health: Explore the impact of work-related stress, burnout, and organizational culture on mental health.
Intersectionality: Examine how factors like race, gender, and socioeconomic status intersect with mental health experiences.
Sometimes the hardest part of research is choosing the right keywords to use when searching for sources. A good set of keywords can be very impactful in obtaining great articles or books on your topic. Generative AI can help with that.
Example: If you are considering research on the topic of honeybee population decline in the U.S., your obvious initial keywords might be "honeybee," "population," or even "apis" which is the scientific word for a honeybee. However, this might not provide you with the more in-depth sources you'd like. Using Generative AI may help to give you more specific keywords to explore.
Copilot provided the following response when asked for keywords on this topic:
Certainly! Here are some keywords related to the factors affecting the honeybee population in the United States:
Parasites and Pests: Varroa mites, small hive beetles, and Nosema spp.
Pathogens: Diseases like American Foulbrood and Deformed Wing Virus.
Poor Nutrition: Challenges due to monoculture farming and limited diet diversity.
Pesticide Exposure: Sublethal effects of pesticides.
Climate Change: Impact on bee habitats and foraging patterns.
Habitat Loss: Reduction in natural bee-friendly environments.
Improper Apiary Management: Practices affecting bee health.
Competition Among Bee Species: Introduced vs. native bees.
Feel free to explore these areas further in your research!
Generative AI can be a useful took to help you study. For instance, it can create a daily study plan for your upcoming exam, or create a guide and flashcards using the information you need to study. The more information you provide to it, the more detailed your study tools will be.
Example: When asked to create a study plan for an exam on Shakespeare's Othello, with only one hour of study for a week, Perplexity provided the following:
Certainly! Here is a condensed outline for a week-long study plan for your exam on Shakespeare's "Othello" with 1 hour of study per day:
Day 1: Introduction
Day 2: Act 1
Day 3: Acts 2 and 3
Plot development, character analysis, important quotes.
Day 4: Act 4
Day 5: Act 5
Day 6: Themes and Symbols
Day 7: Revision and Practice
This structured plan will help you prepare effectively for your exam on "Othello" within a week. Good luck with your studies!
When you're newer to academic research, understanding more complex sources may not come easy. Generative AI may be useful in summarizing or evaluating your sources. For instance, legal jargon can sometimes be confusing. Asking Generative AI tools to summarize a law or legal code could be useful and give you the tools to understand similar language in the future.
Important note: Always check the work of Generative AI. Some tools may not have access to the sources you need summarized and can hallucinate information.
Generative AI can be a useful tool in obtaining quick background information on a subject. It can act as a conversational tool to ask general information questions and obtain details that you do not necessarily need scholarly sources to answer.
For instance, maybe you need a quick answer about the lifecycle of a certain crop, or historical information on a specific country. Generative AI can be a quick way to obtain that information. Though you may also want to consider an online encyclopedia or other reference source to check this information.
Important note: Some Generative AI tools do not have the most current data to generate a response. For that reason, it may be important to fact check the work of AI before using the information in a research assignment.
Generative AI tools go through periodic trainings, or updates, to improve their data output. These tools do not automatically update their knowledge in real-time, and can be up to years behind on the information it knows. For instance, Chat GPT 3.5 is only current up to January 2022. It would not be able to answer questions about events past this date. Some paid version of AI tools may be more up-to-date than others, depending on their training data.
Because of this, the information provided by Generative AI may not be the most accurate and it is imperative to fact-check the information you receive before using it in any of your classwork.
AI hallucinations are fabricated, or made-up, information that is presented as if it were factual. These may be nonsensical or completely inaccurate. The term is used in comparison to human brain activity and perceptions. For instance, humans can sometimes see or perceive images in ink splatter patterns. AI can similarly misinterpret data, or use its imagination, to create inaccurate responses.
This can often happen when you ask AI to provide citations or academic sources. As a result, it is important to fact-check any citations or sources provided by Generative AI tools.
AI bias refers to the result when AI responses reflect human biases within the our society, which can include both historical and current inequalities in the world. This can often occur when the data AI is trained with does not represent reality due to being incomplete or biased.
Analysis has shown that AI image generators tend to favor lighter skin individuals, in that they were portrayed more frequently in generated images of high payed employees, while darker skinned individuals were portrayed in more low paying, labor intensive professions. Similar analysis has been done in terms of bias based on gender.
Nicoletti, L., & Equality, D. B. T. +. (2023, August 1). Humans Are Biased. Generative AI Is Even Worse. Bloomberg.com. https://www.bloomberg.com/graphics/2023-generative-ai-bias/
While Generative AI can create new content such as images, music, and text, it lacks the human creativity produced by artists. This stems from the fact that AI cannot simulate human emotion, feel empathy, or have a true worldview. AI's version of "creativity" mimics the styles in content created by humans, and can only use the data it has been trained on. Therefore, it's content often lacks the emotional depth that comes through human creativity.
When asked about the limitations of generative AI, Chat GPT responded with the following:
