In recent times, many school districts in Bucks County and across Pennsylvania have started to integrate various forms of Artificial Intelligence (AI) training, policies, and instruction. However, AI remains a complex technology that is not entirely understood by everyone involved.
As stakeholders—parents, students, and taxpayers alike—it is essential to assess the quality of your school district’s AI policy.
Here are some crucial questions to raise concerning local AI policies and training measures.
Where is the training sourced from?
Tech companies have invested significant resources in developing their AI technologies, and it is no surprise that they are also pouring money into providing training for K-12 schools, a sector that represents a vast market opportunity.
It is vital to recognize that these corporations are not offering this training as a public service. Google, for instance, has been actively trying to establish its presence in education. An internal document leaked recently indicated that their goal is to create a “pipeline of future users” by embedding itself in educational institutions.
Programs that do not stem from corporate interests, such as the University of Pennsylvania’s Pioneering ASI in School Systems (PASS), provide more reliable options for public school districts. However, caution is warranted as even PASS has received monetary backing from Google.
Corporate sponsorship of AI training is not merely a tactic to foster brand loyalty. Big Tech discussions around AI often presuppose its adoption, focusing on how you will utilize AI rather than whether it should be used at all.
Are ethics being adequately considered?
The topic of AI “ethics” frequently comes up in discussions surrounding AI policies. According to a survey by the Bucks County Beacon, ethics was identified as one of the critical subjects within such policies. However, this is a complicated issue with various facets to explore.
Two prominent ethical concerns within the AI domain are the use of copyrighted materials and environmental impacts.
Companies that develop major AI systems have put forth numerous arguments for why they do not wish to compensate authors for the copyrighted materials used in their training. Ultimately, AI continues to profit from the unpaid work of these creators.
The second concern is environmental in nature.
Data centers have emerged as a pressing issue, where even Elon Musk’s newly launched Grok data center in Memphis operates with unpermitted gas turbines that contribute to noise and pollution. Communities with limited political clout often become targets for such developments. As a result, the escalating demand for power threatens both the advancement of data centers and the stability of the national power grid, with manufacturing slots for gas turbines booked until at least 2030.
In Pennsylvania alone, there are proposals for as many as 50 data centers. Some areas, such as Montour, are resisting zoning changes, while small towns like Archbald, PA face significant challenges, grappling with issues like water and energy depletion, noise pollution, and minimal job creation.
Any program examining the ethical considerations of AI must address these fundamental operational costs.
What capabilities does AI genuinely possess?
Training for students and staff must involve a thorough exploration of what AI can and cannot do effectively.
For instance, if a teacher requests a lesson plan on Hamlet from a chatbot like ChatGPT, it does not analyze various existing lesson plans or determine the most effective pedagogical strategies. It lacks an understanding of Hamlet or the act of teaching in any meaningful sense.
Similarly, when a student employs ChatGPT to compose a paper on Hamlet, the chatbot does not “read” or “interpret” the text accurately.
A study conducted by Carnegie Mellon indicates that AI can lead to greater confidence yet diminished knowledge and critical thinking skills.
Countless individuals mistakenly place their trust in AI’s accuracy, a risky assumption that has become evident in numerous contexts, including the legal field where lawyers have faced dire consequences.
Clarifying how AI functions is inherently complex, but there are accessible resources designed for beginners. One effective summary is that when engaging a chatbot, the inquiry essentially boils down to “What might a response resemble?” Mistakes made by AI—often labeled as “hallucinations”—are not mere glitches; they stem from the fundamental design of AI. In essence, all AI products can be seen as a form of hallucination.
Many people idealize AI as being entirely objective; however, like all computer programs, it reflects the biases ingrained in its data. For instance, a tweaked Grok offering outlandish praise for Elon Musk serves as a humorous reminder that a chatbot can be programmed to reflect the biases of its creators.
Misconceptions abound, with both students and educators perceiving AI as a wise, objective entity. The industry has leveraged these illusions, equating AI to an expert assistant. In contrast, a responsible educational program about AI must ensure participants grasp both its potential and its limitations.
AI is not human
Research shows that interactions between chatbots and individuals—particularly young people—present noteworthy risks. Viewing chatbots as human-like entities can lead to developmental challenges in communication. Since chatbots replicate human language, it may lead many to mistakenly assume there is genuine human-like intellect behind their responses. This is not the case; it is crucial for students to receive consistent reinforcement of this fact.
When is AI acceptable to use?
Several frameworks exist to help students understand appropriate levels of AI use. For example, the Central Bucks’ red zone vs. green zone model is practical, as is the MIT model currently utilized by Council Rock.
School districts must clarify this issue, as there is growing evidence that students overly reliant on AI experience diminished educational outcomes. A Carnegie Mellon study indicates that AI fosters overconfidence while reducing actual knowledge and critical thinking abilities.
Given this context, districts should enforce limitations on teachers’ and staff’s use of AI. If teachers frequently resort to AI for shortcuts, students may internalize the idea that these shortcuts are acceptable.
Who holds accountability?
Be cautious of the notion of a “human in the loop.” Dan Davies succinctly coined the term “accountability sink” in 2024. This refers to the individual within a system who bears responsibility for errors, and in many AI systems, this accountability falls on a human. If a school implements an AI-driven grading system, the person responsible for verifying the AI’s accuracy becomes the accountability sink. This creates a scenario where failure of the AI is perceived as the human’s fault.
If students contravene your established AI usage policy, who takes the responsibility for enforcement? Transitioning from one workload to an alternate responsibility—monitoring student work for AI breaches—does not alleviate teachers’ time constraints.
How is data safety ensured?
Inevitability, the future, and skepticism toward AI
It’s a common trend for tech firms to promote their products as essential, often leading schools to regret hasty decisions. Over a decade ago, schools widely believed that computers were indispensable for education, leading districts to invest heavily in initiatives that provided every student with a personal screen. Today, many are reassessing that choice, attributing declines in test scores partly to the overuse of screens in educational settings and deeming them “mostly useless” .
Tech leaders frequently claim that revolutionary technologies are just a step away from fruition (fully self-driving cars have been “one year away” for over a decade). Such forecasts should be seen more as promotional strategies than genuine predictions.
The expectation that AI will dominate the future must be assessed critically, particularly given the significant sums invested by companies eager to see a return on their capital without materializing into a bubble.
Do today’s students need to learn to utilize AI to thrive in the future job market? Possibly. However, it is equally vital to equip them with a healthy skepticism towards AI. A comprehensive AI “literacy” curriculum must emphasize that AI is not objective, humane, friendly, environmentally sustainable, or an infallible source of wisdom.
A robust AI curriculum in schools should illuminate how AI operates, its applications, and the rationale for its use. Equally important is educating students on when AI should not be employed, fostering an understanding of its limitations and ethical implications.
