With each piece of technology beyond those that simply record and reproduce your own mental process, a student is faced with the question, “is this acceptable?” Certain types of calculator are acceptable in mathematics courses; others are not. An electronic dictionary is acceptable in a language course, but an electronic phrasebook is likely not. An open repository of computer code may offer useful snippets to use legitimately (with citation), but the same source may also offer an entire solution to a problem that you have been asked to solve through your own efforts.
The onset of AI-assisted writing adds a new dimension to the question. Microsoft Word and Outlook will offer to complete your sentences with the most likely subsequent words, and will offer to repair your lapses in grammar. Online researchers will find searches auto-completing or returning AI-generated summaries of the most popular relevant sites. Services now exist that promise to produce entire pieces of academic writing, given a minimal prompt.
|The student operating with integrity will not hesitate to ask a professor if a certain tool is acceptable.|
In every case, it is the student’s responsibility to make sure that the technology they are using is not prohibited. The student operating with integrity will not hesitate to ask a professor if a certain tool is acceptable.
In deciding what tools to use, there are several points to keep in mind:
- Technological assistance is detectable.
Whatever one service offers, another will expose. AI-generated text can be detected. Even more often, the simple contrast between in-class work and work completed with technological assistance is enough to reveal its use.
Your professors are also familiar with the sites that offer answers to assigned problems online. They can find the same material you can, and they likely have seen other students use it.
- The products of technological aides are of dubious quality.
AI text generators, although new, are already notoriously terrible writers. They lack the ability to determine if a piece of text or code makes sense in context. (How often did you really mean to text “ducking”?) Worse, by predicting the next most-likely word, they fabricate “facts” and references that do not exist in reality.
Turning AI-generated text into good text often takes as much work as writing entirely original sentences (or code). Likewise, machine translation cannot comprehend ambiguity, poetry, or any level of contextual meaning; it will render ridiculous errors that take as much effort to repair as writing original work would have taken.
Students may feel like they need the machine’s help to produce “academic” English, but that is untrue. Simple English is good English.
- When you can use it, you’ll have to give credit.
Eventually, more professors will begin to integrate the use of machine-generated text into assignments. As that happens, you will always be required to disclose the nature and extent of AI use in your work, and the failure to disclose it will be counted an act of dishonesty.
- Using technology to help you understand something is always acceptable.
A complex text can be simplified by AI (at the cost of its detailed complexity). A machine translation into your own first language can make a difficult reading assignment manageable (at the cost of vocabulary that may appear on a test). As a first step, both of these are useful techniques. There is no rule that would prevent a student using any tool that may aid in simply understanding; there are only trade-offs that everyone must evaluate for themselves.
AI text is ...