The current development of artificial intelligence is rapid, and the evaluation from all walks of life is also mixed. But how much of this is hype? What can we do to get the most out of AI and other digital technologies in critical areas like education, while ensuring we don't lose sight of core values like fairness, equality, and privacy?
Recently, participated in an open online panel discussion at the Oxford Forum on how to maintain the integrity and quality of education in the face of rapid advances in artificial intelligence and digital technologies.
There is no denying that artificial intelligence will play an increasing role in schools, but many of the more extreme claims go too far. In fact, we don't yet know how this will play out.
But lessons should be learned from existing technology use in the classroom. Like innovations before it, AI and other digital technologies cannot be ignored, but should serve the interests of students, schools, and teachers. Teachers need to be equipped to safely and equitably leverage AI to the fullest benefit of students' education and preparation for our increasingly AI-enhanced world.
Solve the problem
The focus on new technologies in education is not unusual. Previous generations worried about learners using pocket calculators or accessing the free and open internet through Google.
We now know that with the right tools and skills, these emerging technologies can deliver real benefits, with broad implications for how and what we learn. Today, we're used to using calculators and acknowledging that students don't necessarily need to do complex mental math, but instead can focus on understanding how to apply those calculations to solve problems.
Access to Google and other search engines means we no longer have to remember leaders' special dates. But we do need to know how to find accurate sources of information online and develop the critical skills of evaluating that information.
That said, AI poses new risks. Given the impact of education on children's life chances, education should be considered a high-risk environment that needs to be protected to ensure that fundamental rights and vital interests are respected. There are significant ethical and logistical challenges to rolling out AI technologies in schools, including concerns about potential bias, invasion of privacy, accuracy, and effectiveness. With this in mind, and as my colleagues at the Oxford Internet Institute wrote earlier this month, we now need a proactive approach to regulation to ensure we can use these emerging technologies safely across society and in a consistent manner. and do so in a fair manner.
As Rebecca Eynon and coauthors suggest, we need to categorize the different uses of AI and assess the risks and opportunities involved. For example, large language models such as Chat-GPT may perpetuate, or even exacerbate, existing inequalities. At OII, our students and staff are investigating how to make these LLMs more inclusive, such as allowing personalization to reflect the values or needs of different users.
embrace the new reality
Being aware of potential risks doesn't mean we should shy away from technological innovation. We must accept the fact that children grow up in a world where artificial intelligence will become more and more pervasive, and prepare them accordingly. If we fail to teach our students how to learn and use AI, we risk widening the digital divide between those who have access to and can learn AI at home and those who may not have a reliable internet connection at all. The distance learning experience during the pandemic should have reminded us how harmful this disparity still exists.
So, how can teachers and learners be supported to get the most out of AI while ensuring fairness, quality and integrity? The most important factor is that these new tools are thoughtfully adopted and used. The language and imagery used in marketing edtech products can make them sound like they have the answer to every educational problem, but the reality is, few such panaceas exist. However, with careful use, there are still some promising areas of exploration.
Given the ongoing dissatisfaction with teachers' workloads, there is a clear desire to use AI tools for time-consuming tasks such as grading assignments, reviewing grades or checking for plagiarism. Google's Future of Education report highlights AI's potential to save teachers time in the face of teacher shortages, ideally providing more tutoring, guidance and advice to students. Of course, this only works if the school refuses to increase class sizes in response.
The potential for resource personalization is obviously interesting as well. Where tests or assignments are completed using apps or online courses, learning technology can provide instant feedback to help learners identify areas they need to focus on. In this context, it is crucial to design AI-based educational tools together with educational experts if high-quality outcomes are to be ensured.
Finally, students themselves must learn how to use innovative AI tools like ChatGPT, and how to use them responsibly and critically. Just as calculators have changed the way we do math, large language models may change the way we write and present knowledge. Schools and universities play a central role in setting these new norms and expectations.
Think of AI as a tool in your toolset. We need it to improve education through existing experience, and to do this we need to take the time to reflect and regulate its use, avoiding unforeseen incidents.