Two Stories of Artificial Intelligence

According to Oxford Languages, artificial intelligence is “the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages” (Oxford English Dictionary). To provide a simple summary, artificial intelligence is man-made intelligence. However, the narrative surrounding this technology is anything but simple, divided into two stories with entirely opposing perspectives. One, told by proponents of AI technology, paints a picture of sweeping technological advancements beyond current conception—advancements that can provide a multitude of extraordinary benefits to human society. The other, told by opponents of AI, focuses on job elimination, tales of cheating with Chat-GPT, and possible future destruction. One side’s story is overly optimistic; the other’s, Terminator-esque. However, neither those who support AI nor those who oppose it are entirely right or wrong. The true answer lies somewhere in between these two extremes, using caution to guide the development of new artificial intelligence technologies. 

‍

One field that especially requires a balanced solution to artificial intelligence's growing influence is education. With the growth of chatbots like Chat-GPT and the increased array of possible AI education tools for teachers, AI already has a foothold in the classroom, which will only continue to grow. Thus, it is essential for schools to use AI in a way that provides both teachers and students with more support while preventing harmful side effects. 

‍

Artificial Intelligence for Students

In the discussion about artificial intelligence in education, policymakers must first consider how students can use artificial intelligence technology. The discussion about student use of AI can only occur with mentioning Chat-GPT, a current topic of controversy, and other chatbots. Chat-GPT, “a large language model” (Sundar 2023), is an artificial-intelligence-powered chatbot that provides students with an accessible way to cheat. It is currently extremely popular; though there are many other chatbots on the market, Chat-GPT is the most well-known. Chat-GPT can easily write a detailed essay about any topic a student chooses, and teachers often struggle to separate Chat-GPT-authored essays from student-authored essays. Students can use chatbots to draft, edit, and create assignments nearly identical to original work. According to a study conducted by Common Sense Media, the majority of teenage students have used Chat-GPT, making the chatbot a popular tool for students and a true concern for teachers. Furthermore, a Forbes poll found that 1 in 5 college students have used AI on homework, proving that AI usage is prevalent through all levels of education. 

‍

Artificial Intelligence for Teachers

On the other hand, teachers can use artificial intelligence for greater teaching assistance. One of the main ways teachers can do this is through using “adaptive learning platforms'' (Akgun and Greenhow 2021), which tailor students’ “learning materials” (Akgun and Greenhow 2021) to their current progress in classes. In particular, these learning platforms can combat the issues resulting from larger class sizes, allowing students in larger schools to experience a quality of education similar to that in a smaller school. With artificial intelligence-powered educational technologies, teachers can provide more flexible learning plans. 

‍

With artificial intelligence, teachers can not only provide detailed learning support, but also provide more emotional help in the classroom. Using “facial recognition software and predictive analytics” (Akgun and Greenhow 2021), teachers can easily observe students’ emotional and behavioral state in class, which can show them areas where they may need to make a lesson more engaging or easier to understand (Akgun and Greenhow 2021). Additionally, predictive analytics can “detect patterns about learners” (Akgun and Greenhow 2021), allowing teachers to see when a student may need additional help with their assignments. Furthermore, teachers can use artificial intelligence to alleviate students’ negative emotions surrounding asking for help. With artificial intelligence programs, students can receive “constructive feedback” (Chen 2023) without the “fear of judgment” (Chen 2023) that comes with finding help from others. As students feel more empowered to seek help, they will seek help more often, and they can gain a better understanding of their assignments and likely improve their academic performance. 

‍

Furthermore, according to a report from the National Institute of Health, teachers can use AI to “provid[e] students with detailed and timely feedback on their writing products” (Akgun and Greenhow 2021). Instead of having to wait for feedback that may be incomplete or rushed, students can receive excellent feedback quickly, which will likely lead to better future assignments. Without any technological assistance, a teacher in a class of about thirty students cannot hope to give students the detailed, individualized feedback they need; however, with AI, the possibility for more comprehensive comments opens up. Teachers can also use artificial intelligence to grade assignments, which will give them more time to prepare classroom materials. Additionally, artificial intelligence programs like chatbots can also “serve as practice students for new teachers” (Chen 2023), allowing teachers to try out different teaching plans and styles to find the plan that will be best for students. AI programs can also offer “post-lesson reports” (Chen 2023) for teachers to use, providing them with even more opportunities to optimize their classroom. Overall, artificial intelligence can support teachers and allow for more individualized learning.

‍

Risks of Artificial Intelligence in Education

While artificial intelligence in education may seem like the perfect solution to the issues that plague schools across the world, providing individualized learning and greater assistance for both teachers and students, it does not come without risks. Free, easily accessible chatbots make cheating uncomplicated, as they quickly produce excellent assignments. If teachers cannot always separate AI assignments from real assignments, cheating can run rampant in schools, preventing students from actually learning the material. Furthermore, if students are overly reliant on AI assistance, then they cannot become independent learners, and they certainly cannot gain the skills necessary to succeed later in life.

‍

Use of AI to fulfill teachers’ roles can also prevent students from learning properly and developing necessary interpersonal skills. According to the Harvard Business Review, “face-to-face experiences inherently have the potential to generate and sustain focus” and human interaction can lead to “deep learning” and “interpersonal bonding” (Hooijberg and Watkins 2021). Suppose students spend a substantial amount of their time in the classroom using artificial intelligence programs like adaptive learning platforms. In that case, they cannot focus as well as if they were talking to a teacher, meaning that their study time will be less productive and efficient. Furthermore, if the majority of students’ time is spent on a computer, they may lose social skills necessary for success in careers and relationships. A conversation with an AI chatbot is a paltry substitute for a conversation with a human being. Artificial intelligence chatbots do not possess humans' emotional abilities, and they certainly cannot dive headfirst into complex philosophical conversations. Thus, a focus on artificial intelligence in schools could prevent “deep learning” (Hooijberg and Watkins 2021), which involves philosophical conversation (Hooijberg and Watkins 2021). Deep learning also emphasizes the ability to “understand … subtle differences in context” (Hooijberg and Watkins 2021), an ability which chatbots simply do not have. If artificial intelligence becomes ubiquitous in schools, students will likely experience less face-to-face interaction with teachers and other students, which can stunt their development.

‍

Additionally, artificial intelligence programs pose risks of bias, as “algorithms … [include] data that represent society’s historical and systemic biases” (Akgun and Greenhow 2021). Thus, such biases can be expressed through the programs students use in class, specifically in areas such as language classes. Additionally, some biases are expressed through programs that consistently present “some groups in a negative light” (Gaskins 2022). Children who belong to communities that are portrayed negatively may develop self-esteem and mental health problems, especially if they are exposed to such biased portrayals on a daily basis. Rather than supporting students’ learning and development, the bias in artificial intelligence could harm students’ development. Additionally, even the use of facial recognition technology contributes to biases. According to a paper by Nettrice Gaskins, “facial recognition software” (Gaskins 2022) in some schools “has problems identifying the faces of African-American students” (Gaskins 2022). Schools that use artificial intelligence-powered facial recognition technology could potentially exclude many of their students from reaping the benefits of such technology simply because of ingrained bias against students with darker skin colors. 

‍

As well as perpetuating biases, artificial intelligence in schools raises a variety of questions about privacy issues. Using facial recognition technology to constantly monitor students’ emotional state and collect that data clearly violates their privacy. Additionally, even AI-powered learning software uses “tracking systems” that “not only necessitate monitoring of activities but also determine the future preferences and actions of their users” (Akgun and Greenhow 2021). Tracking and monitoring the behavior of students both in the real and virtual world seems like a scenario from one of Orwell’s novels. Furthermore, in many cases, there is no way for teachers and students to be sure that artificial intelligence companies have truly secured their data, creating even more reasons for concern.

‍

Finally, costs must also be taken into consideration when discussing the risks of implementing artificial intelligence in schools. Olufemi Shonubi, the founder of an education technology company, writes that “the cost of developing and implementing AI-powered educational tools can be a significant barrier for schools and teachers” (Shonubi 2023). Suppose only the wealthiest schools can afford to implement artificial intelligence programs in the classroom. In that case, many schools will be excluded from using such technology, worsening the “performance gaps” (Garcia and Weiss 2017) between well-funded and underfunded schools. Students’ quality of education will suffer simply because of the expenses necessary to use artificial intelligence technology. 

‍

Policy Options

Though using artificial intelligence in schools can provide many benefits to students and teachers, artificial intelligence carries various risks, from privacy violations to excessive costs. Thus, policymakers must approach artificial intelligence with caution, emphasizing teacher and student welfare while minimizing potential drawbacks. Some school districts across the country have already taken swift action against Chat-GPT, with New York City Public School District, Los Angeles Unified School District, and Seattle Public schools, and others blocking it “on all school devices” (Johnson). With Chat-GPT’s prevalence among students and its contribution to cheating, such a policy is wise. Furthermore, the policy does not completely ban students from using Chat-GPT; rather, it only prevents Chat-GPT on school devices, meaning that students can access the program if using a device that is not connected to the school. Students can still use Chat-GPT when they are not on school grounds—this policy simply prevents cheating on school devices, a reasonable step. However, simply banning Chat-GPT alone will not solve the many risks resulting from the use of artificial intelligence in education. As AI becomes both more advanced and more widespread, it is inevitable that schools will begin to incorporate more AI-focused activities. Instead of completely banning artificial intelligence technology in schools, both school districts and states can take steps to implement positive artificial intelligence programs while minimizing any risks.  Leaders must ensure that schools are properly prepared for the implementation of artificial intelligence in the classroom, and multiple policies can achieve this goal. To prevent the harmful developmental effects potentially resulting from artificial intelligence in the classroom, state legislators could introduce bills limiting AI-based activities to a certain percentage of classroom time to ensure that human interaction makes up the majority of class time. Thus, students can experience the benefits of artificial intelligence technology while still participating in enriching discussions and retaining essential learning and communication skills. Furthermore, to combat racism in artificial intelligence, school districts can intentionally choose to partner with “edtech” (Shonubi 2023) companies that focus on eliminating racism in artificial intelligence. School districts can also make sure to partner with companies that will protect their data to prevent any privacy breaches. Additionally, to prevent privacy breaches related to facial recognition technology, states should take steps to prevent the implementation of facial recognition technology. New York’s State Senate recognized this in 2020, passing a bill that “prohibits the use of biometric identifying technology in schools until July 1, 2022” (New York State Senate). The harms of using facial recognition technology far outweigh the benefits, as facial recognition technology stands as a serious threat to privacy. Thus, state legislators must take steps to prevent the use of such technology, following in the footsteps of the New York Senate. Finally, another policy option for school districts to combat possible misuse of artificial intelligence is to mandate artificial intelligence training for both students and teachers. If a school district decides to use artificial intelligence in classrooms, the district should ensure that both teachers and students are properly educated about the risks of AI. With such policies, leaders can introduce the positive aspects of artificial intelligence into the classroom while avoiding negative side effects.

Conclusion

With common-sense policies, schools can smoothly adopt artificial intelligence in the classroom. Rather than leaning into one of the divided stories surrounding artificial intelligence, policymakers can write a new story in which leaders avoid both reckless optimism and extreme pessimism. Rather than allowing emotions to influence decisions, policymakers at the state and district level should analyze the possible consequences of artificial intelligence from a balanced perspective, prioritizing student and teacher safety. Hopefully, with a thoughtful approach, artificial intelligence can add to the classroom experience, helping, not hindering education.