Artificial Intelligence (AI) is the simulation of human intelligence in machines programmed to think like humans and mimic their actions. This field of study combines computer science and robust datasets to enable problem-solving and decision-making capabilities in machines. AI encompasses various subfields, including Machine Learning, Deep Learning, and Generative AI.
 |
 |
 |
 |
1950s–1970s |
1980s–2010s |
2011s–2020s |
Present Day |
| Neural Networks | Deep Learning | Generative AI | |
| The perceptron, an early neural network model, demonstrates the potential for machines to learn and make decisions. | The development and refinement of backpropagation algorithms enable efficient training of multi-layer neural networks. | The advent of transformer models revolutionize natural language processing and establish new frontiers in generative models and broader AI applications. | Integration of multimodal learning capabilities, enabling AI systems to understand, generate, and translate across various forms of data like text, images, and sounds cohesively. |
The use of AI in academic work depends on its role in the scholarly process varies. AI as a grammar enhancer or technical aid is generally accepted, as it improves presentation without changing the main content. Ethical concerns arise when AI starts generating significant content or mimics academic voices. The level of AI's content creation influence is key to its acceptability. Balancing AI assistance with maintaining the author's authenticity is crucial in academia's evolving AI landscape.
Below is a helpful Framework for GenAI Involvement Levels by the University of Michigan (adapted from Piers Steel, University of Calgary) on how to think about the various levels of GenAI involvement in your work, going from most acceptable to most controversial.
This content is offered under a CC Attribution Non-Commercial Share Alike. Content in this box can be considered under this license unless otherwise noted.
Here are 5 general rules that can guide researchers and educators in navigating the complexities of Generative AI:
| Number | Rule | Details |
| 1 | Understand the Black Box Nature of AI |
Generative AI models often function as "black boxes," meaning their internal decision-making processes are not transparent. This can pose challenges for researchers who need to understand how conclusions are reached. When utilizing generative AI:
|
| 2 | Be Cautious of Data Biases |
Generative AI models are trained on large datasets that contain inherent biases. The outputs can reflect these biases, leading to skewed or prejudiced results. To mitigate this:
|
| 3 | Consider the Generalization Capabilities |
AI models can excel in tasks they are trained for but might struggle with generalizing to new, unseen situations. When applying generative AI in research:
|
| 4 | Acknowledge the Ethical Implications |
The use of generative AI raises significant ethical concerns, including privacy issues, the potential for misuse, and the societal impact of decisions made by AI. Researchers should:
AI cannot replicate complex human emotions or ethical reasoning, limiting its applicability in contexts requiring deep empathy or moral judgment. |
| 5 | Stay Informed About AI Developments |
AI is rapidly evolving, with continuous improvements and changes in capabilities, methodologies, and ethical standards. To effectively use generative AI:
|
Acknowledging these limitations is critical for responsibly navigating the development and application of AI technologies. Incorporating these rules of thumb into your research practices with generative AI can help navigate its limitations while harnessing its potential.