To What Extent Should Artificial Intelligence Be Regulated for Schools?

Artificial intelligence (AI) could change the world in unprecedented ways. All of that potential is exciting, but it also raises some concerns — especially when you realize AI use is growing much faster than people’s understanding of it. These concerns lead to questions about its regulation, especially in sensitive areas like schools.

As tools like ChatGPT have led to more AI use in schools, some people wonder if the government should step in to regulate it. If so, how far should these laws go? What should they stop and what should they allow to let students and teachers make the most of this technology?

Deciding how much to regulate AI in schools starts with understanding its benefits and risks. You must look at how schools could use AI, which falls into two basic categories — AI for administration and AI for education.

The first category deals with behind-the-scenes use cases. AI could automate time-consuming work like scheduling, filing paperwork, and grading some tests to give teachers more time with students. Considering how teachers work a whopping 54 hours a week and spend less than half of that time actually teaching, these time-savers could be a huge help.

AI could also analyze student behavior or performance to predict future outcomes. These predictions would let teachers see who may need intervention to ensure they stay in school and learn to the best of their abilities.

These benefits are impressive, but using AI in this way involves feeding a lot of student data to algorithms. That could introduce privacy and cybersecurity concerns. There’s also the issue of bias, which AI tends to exaggerate if its training data contains human biases. If schools don’t address these problems, AI use could worsen inequality in education.

AI could also help teachers offer more engaging lessons. Technology lets students explore historical events from the classroom or gain hands-on virtual experiences they couldn’t get in the real world, and AI takes those further. Students could use it as a tool to make research fun or see how technology changes the world.

AI could also help personalize learning by recognizing which lessons work best with different students. They could then help teachers adjust their classes to individual students’ needs to improve their education. As AI becomes a bigger part of business, learning how to use AI could become an essential part of preparing for a future career.

The danger here is that AI has some accuracy issues. You’ve likely already seen how ChatGPT can spew false or misleading information, so using AI as a teaching tool could teach students things that aren’t true. If these falsehoods contain harmful biases, that could be an even bigger issue.

Given the benefits, AI is too useful for both teachers and students to ban it in schools altogether. At the same time, its risks are too big to let it go completely unregulated. AI in schools needs some restrictions across a few different areas.

First, regulations should require AI in schools to meet strict privacy standards. That starts with training education AI on either synthetic or de-identified data to ensure any issues don’t accidentally reveal students’ real-world information. Education was the most-targeted industry for cyberattacks in 2022, so these protections are crucial.

Regulations should also guide how much information schools can feed into AI tools outside of training. Some data-sharing is unavoidable, especially if schools are using AI to manage student records. In those cases, the law should require schools to scrub student data of identifiers as much as possible and restrict access to AI databases through encryption and zero-trust security.

Educational AI regulations should also address the issue of bias. The key here is to be aware of human biases and how they influence AI models in training. Schools should only be able to use AI tools that have strict bias prevention measures in place.

De-identifying training data is a good start, as it removes factors like race or gender from the AI decision-making process entirely. AI development teams must also have diverse boards to catch and fix potential issues early. In both training and use, schools and dev teams should regularly review AI processes and results for signs of potential bias. Final decisions should always fall to people, not AI.

Finally, AI in schools should meet accuracy-related standards. If teachers rely too heavily on AI that could contain misleading or false information, they could harm students’ education. Consequently, schools should only use AI where the developers take great lengths to ensure the model’s reliability.

Since it’s virtually impossible to guarantee an AI tool is 100% accurate all of the time, the most important side of this issue is addressing how schools use AI. Teachers should always fact-check AI, express to students how it can be inaccurate, and only use it as support, not the main teaching method. 

Schools should use these potential inaccuracies as a learning opportunity. Experts say almost every business will use AI in the next 20 years, so teaching students about where it falls short is crucial for equipping the next generation of workers for success.

AI’s potential in education is too big to ignore. That said, it carries some undeniable risks. Given these pros and cons, school systems and governments should focus on regulating the technology to ensure teachers and students use it safely.

If you can take all these risks into account, AI will revolutionize education for the better. If not, it could have dangerous implications. The solution is not to avoid AI in schools but to approach it carefully and do so before it becomes an unofficial industry standard.