It’s been a tough year for faculty in various disciplines when it comes to artificial intelligence. Many companies have redesigned challenges and developed new course policies in the presence of generative AI tools. Some have pondered at conferences or in their spare time what makes prose human. (One possible answer is burstiness.) Some have designed, conducted, or participated in AI-focused workshops in teaching and learning, with or without support. One sent a message to a student saying, “Don’t grade this chat with her GPT.” (Definitely how long Fallout takes.)
Amidst the chaos of AI in 2023, professors are also grappling with paradoxes. In one story, large-scale language models provide much-needed help by enhancing students’ creativity, research, writing, and problem-solving skills. Students with disabilities also benefit from AI tools that provide executive function support and more. But in another narrative, algorithms have the potential to reproduce systemic biases and widen educational inequalities like technology never existed before.
“There are two schools of thought,” said Berkassem Karim Bugida, director of the library at Stony Brook University, adding that he emphasized the middle ground. “All data is biased, allowing us to approach our tools in an ethical and responsible manner,” he said.
AI literacy gap
Some students already have advanced skills in prompt engineering, the art of creating questions for natural language processing tools for better results. Some people have little experience in talking to machines.
“What’s happening is that the rich are getting richer, so to speak,” says Louis Ludwig, a mathematics professor and director of the Center for Education and Learning at Denison University. “People who know what they’re doing can actually get this to sing, but people who don’t know how to use it are kind of left behind.”
Many instructors are working to help their students overcome this AI divide. But such efforts require fine tuning. In an ideal world, a student would reach a rapid engineering sweet spot, where she could leverage her AI tools for learning without hindering her personal or academic growth.
According to Laura Domin, an English professor and director of the Technical Writing Program at the University of Central Oklahoma, students are sometimes more confident in the output of AI tools than in their own work. As a result, some people may not put much effort into shaping the output of AI tools.
When ChatGPT launched in late 2022, many students were halfway through the academic year at a familiar institution. However, students will be redistributed at the start of the new semester this fall. Some may join the company with detailed knowledge of AI tools. According to a 2023 report, other primarily rural states in the United States, such as West Virginia, Alaska, Mississippi and Arkansas, have poor internet quality, latency and access in their geographic areas. may have limited digital literacy.
Even someone from a state that ranks highly for broadband can be at a digital disadvantage. For example, California ranks high, but her 40% of Latino students in the state do not have reliable broadband access to her.
And some freshmen may have graduated from high schools that banned AI-powered tools. Some have extensive teaching experience with technology. In Australia, for example, ChatGPT is banned in most public schools, raising concerns that it could create an information divide between students in the country’s public and private schools. Even in the US, many students have ready access to tools like her ChatGPT on their mobile phones in their pockets, but some students don’t own laptops or mobile phones.
“We will see a stratification of who already knows how to circumvent AI, who doesn’t, who knows more about the dark side and who is entering for the first time,” Dumin said. “That’s a big concern for stocks.”
Paid AI tools
“The most robust AI is behind paywalls,” says Emily Isaacs, Executive Director of the Office of Faculty Advancement and Professor of Writing at Montclair State University. inside higher education. At the University of Montclair, nearly half of the students are recipients of Pell grants, which indicate low-income status, and nearly half of the faculty are temporary employees, Isaacs said. “Our faculty and students have financial stressors. [AI tool] should i buy something? 』
Students’ unequal access to premium versions of educational technology reflects the inequalities Isaacs (in her words) observes daily in commuter schools.
“Students with cars can get to campus much easier than those who take three buses from their home in Newark to campus,” said Isaacs. To promote equity, Isaacs wants his AI-powered educational products to be offered as open educational resources. Until that day comes, she wonders whether that cost can be folded into tuition and fees in the same way that, say, all students at Montclair State University are given access to other proprietary software. I am thinking.
In the absence of equity-focused policies, some academics recognize the role their units and departments may play in their future career paths.
“Academic libraries are likely to suffer further downturns in the future,” Bugida said. “It would be the same as managing the licenses of journals … Some universities will not have the means to subscribe. So there will be a digital divide.”
Isaacs said professors who encourage the use of AI tools should be mindful of the costs they impose on students until universities develop policies to address inequalities.
culturally insensitive chatbots
Colin Björk, a senior lecturer at New Zealand’s Massey University, recently called non-English languages, including oral and indigenous languages, “edge cases,” a term used to describe unusual cases that confuse computers with Microsoft executives. talked. That’s because large-scale language models are trained online, and the dataset is often standard American English. Because of this, AI output often does not represent the depth and breadth of many students’ multicultural and multilingual experiences.
“We try to teach our students to find their own voice,” Dumin said. As an example, she said, African-American native English and Appalachian English-speaking students might find generic texts generated by AI to be more effective than their own. “We may lose some of the diversity of letters and sounds. It would be really sad.”
“Black English is important,” JSTOR Daily resident linguist Chi Luu recently wrote. Lu argues for “almost constant creativity” in linguistic innovation while embracing rich regional and class differences. Its innovations told stories of migration and movement, leaving its mark on Standard English by “sliding seamlessly into the language of art, music, poetry, storytelling and social media.” But that feeling of alienation can sometimes have a negative impact on those who speak the language, such as in job interviews, renting an apartment, or dealing with the police, Lu wrote.
Many academics are concerned about training large-scale language models on the Internet, where biases are established.
“There will be a normalized expectation that everyone else will have to change in order to come close to the default of this machine’s generative AI language bias,” he said, targeting adult learners facing critical challenges. said Lance Eaton, Director of Digital Education at College Unbound, a higher education institution. barriers to entering university. College Unbound’s bachelor’s degree programs are designed around a personalized, interest- and project-based curriculum.
empowering chatbots
Language learners can use AI writing tools to learn vocabulary, genres, idioms, grammar, and more. Likewise, those who struggle in social settings due to neurodivergence or who fear being judged by their peers may benefit. Chat bots are skilled and willing conversational partners, and conversing with chat bots provides a low-risk way to experiment. It can in turn support learners’ self-confidence.
“Being able to ask anything is really powerful,” Eaton said. “I don’t feel any judgment from the computer.”
Practicing conversational skills improves social mobility, especially for people with communication disorders. For example, Fiona Given, a lawyer who lives with cerebral palsy, often used assistive technology before ChatGPT to save money when writing messages, according to the article. conversation. Still, she worried that her minimalist response would be perceived as “curt, if not rude.” Given has found that ChatGPT helps add the politeness to an email, saving time and showing professionalism.
AI-assisted Career Advancement
Many employers require job seekers to write and submit a cover letter when applying for a job. This is often true even if the job skills required do not overlap with the skills of writing a cover letter.
“The way we express ourselves, the language we use, the nuances we are expected to use… [most] That time has nothing to do with the actual job you’re applying for,” Denison’s Ludwig said.
During his 21 years at the Ohio Liberal Arts College, Ludwig has served on numerous faculty recruitment committees, especially in the mathematics department. During that time, he and his colleagues have placed great importance on applicants’ cover letters.
“Perhaps we have done something wrong,” Ludwig said. “People who are simply not good at writing or who are irrelevant will not pass.”
Siladitya (Raj) Chowdhury, executive director of the Center for Innovation Learning at the University of South Alabama, also participated in the panel and recognized biases among evaluators that were partly caused by language.
AI writing tools “help mitigate the impact of someone expressing something, even if the content is good, if the means of language for that particular purpose can be standardized,” Chowdhury said. said.
Despite this, many employers continue to require cover letters for open positions. Her AI assistance in job applications “could pave the way for people who were previously unable to even get an interview,” Dumin said.
But opinions among HR professionals are divided, and about the use of AI assistance in job searches. Some see it as a “marketable skill”, while others see it as a “dealer.”
AI time-saving features
Few university professors have training in education, but some do. New faculty with minimal teaching experience can benefit from generative AI tools.
“This could make what the new teacher’s lesson plans look like and more powerful,” Dumin said. “When you have a solid lesson plan and it builds on what you feel is good teaching, you can come into class with a little more confidence.”
Dumin has seen AI-generated lesson plans, including feedback from colleagues who trust the premium version of ChatGPT. They report that it is “much better than the free version”. Douming said he was “very optimistic” about the outcome and was “not unimpressed”. The term makes her laugh and alludes to the ambiguity surrounding her conversations about AI in her education.
For students and teachers struggling with executive skills or lacking time, AI tools can help prioritize tasks, organize information, and create schedules. According to a World Economic Forum report, AI could free up 20 to 30 percent of faculty and staff time from mundane administrative tasks and reallocate it to activities that support student learning. It can make a noticeable difference in an area where practitioners cite fatigue and burnout from unmanageable administrative tasks.
Meanwhile, educators and policy makers are making cautious efforts aimed at maximizing fairness and minimizing bias in AI. The White House Office of Science and Technology Policy has released a blueprint for the 2022 AI Bill of Rights, envisioning a future where stock indices are embedded in algorithms.
And last month, the U.S. Department of Education’s Office of Educational Technology released a report that provides insights and recommendations on AI in teaching and learning. The report is 71 pages long, but if you press it, the message may be summarized in the five words contained within. “emphasizing human involvement”;
