The Implications of AI Writing in Education and Society

Image Source: Unsplash

When OpenAI released its chatbot, ChatGPT, last year, many AI enthusiasts feared the worst for fields like screenwriting, computer programming, and music composition. The education industry, in particular, immediately experienced the power of ChatGPT. Students could easily cheat on essays and college admissions applications, while teachers could assign heavy lesson preparations to AI without raising any suspicion.

However, ChatGPT is not the final say in education. As soon as students started using chatbot conversations as their own work, a new wave of AI programs emerged to detect AI-generated writing. Teachers began to incorporate ChatGPT responses into their lesson plans to keep up with the learning trend.

In reality, AI has the potential to enhance critical thinking and expand soft skills among students. Yet, there is concern that children might become too reliant on AI to answer questions, leading to a decline in basic skills and knowledge. Psychologists Edward Deacy and Richard Ryan propose that humans are driven by autonomy, relevance, and competence. Therefore, students will continue to learn regardless of any shortcuts that emerge. The creation of the encyclopedia is an example of this. Students do not abandon their studies of arts and sciences because they can quickly search the web for dates and formulas. Instead, they can use web tools to test their answers, which aids in their learning progress.

Since education is one of the first consumer use cases for AI, programs such as ChatGPT will expose millions of children, teachers, and administrators to AI technology. Therefore, it is necessary to focus on the applications of AI and its impact on our lives. In the following paragraphs, we will explore five predictions for AI and the future of learning, knowledge, and education.

I. The mainstreaming of one-to-one models

In the past, one-on-one models like coaching, mentoring, tutoring services, and therapy were only available to the wealthy. However, with the help of artificial intelligence, these services can now be democratized and made available to more people. Bloom's "2 Sigma problem" discovered that students receiving one-on-one instruction performed two standard deviations higher than children in traditional classrooms, and AI can be used to address this issue by providing personalized learning plans and emotional and behavioral support. For example, Numerade's AI tutor, Ace, generates personalized learning plans that curate appropriate content based on a student's skill level.

AI can also make experts and academic luminaries accessible to all learners, breaking down resource barriers. This development has significant democratizing implications for professions where mentorship and apprenticeship are critical. For instance, Delphi is trying to create a platform where a startup founder can chat with an AI version of Marc Andreessen or Paul Graham at any time. Similarly, apps like Historical Figures and Character AI allow users to talk to important historical figures or create "characters" to talk to.

In fields susceptible to stigma like mental health, AI-enhanced solutions such as Replika or Link may be more accessible than human therapists. They are cheaper and available 24 hours a day, encouraging patients who are afraid to be seen by strangers. Additionally, AI can be personalized and tailored to individual style preferences, solving a problem in the psychotherapy industry that is difficult to identify and match. The marginal cost of augmented AI therapy software is low, allowing for more affordable end products that provide mass-market access. However, AI is not yet capable of 100% human-level thoughtfulness and expertise, and there are times when people may only want to interact with humans in real life.

II. Personalized learning becomes achievable with AI

Artificial intelligence has the potential to bring personalized learning to a whole new level.

Through AI, we can personalize the learning mode and various aspects of the content, including visuals, text, and audio. Additionally, we can tailor course content based on a user's preferences, such as their favorite character, hobby, or genre. The software can track a user's level of knowledge and test their learning progress, and customize content to address any gaps and better understand their learning goals.

For instance, Cameo has introduced a children's product featuring popular IP characters like Blippi and Spider-Man. One mother asked Spider-Man to encourage her child to use the toilet, and it seems to have worked wonders. Moreover, AI can cater to different types of learners, including those who are advanced, falling behind in certain subjects, hesitant to speak up in class, or have special learning needs.

III. A new generation of AI tools for teachers and students may take the lead

Image Source: Unsplash

The next generation of AI tools is likely to take the lead in education, with teachers and students as early adopters. Historically, educators and students have played a crucial role in productivity software, providing feedback to developers and founders to optimize their products. This trend applies to chat-based conversational interfaces and next-gen AI-based tools, which are expected to be popular among teachers and students.

Moreover, teachers are often overworked and underfunded, leaving them with little time to focus on students. They spend most of their time grading, creating lesson plans, and preparing for classes. AI can reduce their workload by learning from millions of educational materials and generating personalized output for their individual classrooms, freeing up time for teachers to give personalized attention to their students.

Similarly, students are always looking for ways to save time and improve their performance. Chegg has been a popular resource for the previous generation, but now new AI-based resources like Photomath and Numerade can help students solve complex math and science problems. These tools can spread quickly through universities, student organizations, and social clubs, making them a popular choice for students looking for creative ways to save time and get ahead.

IV. AI writing is included in the evaluation system and new assessment rules are developed

The use of AI-assisted writing tools, such as ChatGPT, has sparked debates in public education circles. Some schools, like those in New York and Seattle, have temporarily banned the use of such tools, including college admissions applications. However, many educators argue that AI is a critical career skill for the future and should be integrated into learning and teaching. To achieve this, classrooms need to adapt and assessment methods must be developed to measure student outcomes. Just as calculators and personal laptops became key classroom technologies, a new generation of AI-based tools is emerging that can better assess and certify student learning outcomes.

However, we must consider that the availability and accessibility of these technologies may create disparities among students. Those without access to AI tools due to school policies or lack of resources may be at a disadvantage compared to their peers who can use them at home. This disparity could also widen the gap between private and public schooling, as private schools are more likely to have the resources to adopt and integrate new technologies. Therefore, it is important to develop fair and equitable policies that ensure access to AI-based tools for all students regardless of their background or resources.

V. AI may produce misinformation and bias, it is crucial for us to conduct fact-checking

Algorithms rely on human behavior and judgment, which means that biases based on social factors such as race and gender are embedded in them, and these biases will continue to be amplified. For instance, Gmail's AI assumes that investors are male. Despite Google's attempts to fix the problem, it remains unresolved. In this biased environment, fact-checking becomes crucial, especially when AI produces misinformation. AI-generated responses are dangerous because they can create a false sense of accuracy and credibility. A study conducted by the University of Washington showed that 72% of people reading an AI-generated news article believed it was credible, even though it was factually incorrect.

As AI-generated content becomes more prevalent, it is essential to filter for quality and factual accuracy. The mainstream public will likely trust branded media sources more than user-generated content. However, blind trust in certain brands, experts, or individuals can be problematic. A lack of basic understanding of details could lead to difficulties in edge situations and crises. For example, as web development becomes more abstract, front-end engineers may have little exposure to backends and databases. While this abstraction empowers less-skilled users, a critical failure on the backend could become catastrophic if no one understands how to fix it.

In conclusion, as AI continues to evolve, it is important to address the biases that exist in current algorithms and conduct fact-checking to combat misinformation. Additionally, we need to be aware of the potential consequences of creating a generation of people who lack a basic understanding of details, as this could lead to difficulties in crisis situations.