Proponents of Artificial Intelligence (AI) argue that it holds great promise for increasing equity in education by personalising learning experiences, providing immediate, individual feedback, and identifying and addressing individual student needs. But AI also has the potential to exacerbate inequity by widening the digital divide, increasing digital distractions and introducing bias into pedagogy.
The PISA 2022 results provide a critical lens through which we can examine disparities in education and highlight the urgent need for reforms, especially as we navigate the burgeoning AI era. By understanding the implications of these results, we can work towards a more equitable educational landscape, ensuring that advancements in technology benefit all students, regardless of their socio-economic background.
Ensuring equitable access to AI tools is crucial. PISA 2022 found a negative association between a lack of or inadequate digital resources and student performance. High performing education systems ensure that every student has access to a digital device and high performing schools, which tend to have a more socioeconomically advantaged student body, suffer less from shortages of digital resources. Disparities in the shortage of digital resources were also observed between rural and urban schools (in 21 education systems, rural schools suffered more from shortages) and between public and private schools.
Yet it is important to note that the availability of computers does not in itself indicate a school’s preparedness for digital learning. Having adequate guidelines and support structures for their use is also important. On average across the OECD, the biggest improvements in schools’ preparedness for digital learning observed between 2018 and 2022 concern the availability of an effective online learning-support platform, teachers having the necessary technical and pedagogical skills to integrate digital devices into their instruction and the availability of effective professional resources for teachers to learn how to use digital devices.
While a majority of students across the OECD feel prepared to use digital technology for learning, a significant majority remain digitally excluded. Not surprisingly, those from socioeconomically disadvantaged backgrounds or from rural areas fare the worst. In Australia, for instance, nearly a quarter of Australians are digitally excluded in 2024 (Bentley & Naughtin, 2024). Not having adequate access to computers and the internet reduces familiarity with digital technology, erodes confidence and fuels disengagement.
This has implications for AI. Those that are less confident in digital technology are likely to feel less positively about AI. This matters because, as OECD (2023a) research has shown, AI confers significant benefits to workers able to leverage it in the workplace. As the AI revolution accelerates, those workers able to use AI effectively will have a growing advantage over those who do not.
This matters in the classroom as well. As schools move to adopt more AI tools for learning, we must equip all students with the knowledge, behaviours and tools necessary to navigate them successfully. Otherwise, we risk depriving young people of access to a new generation of learning tools.
Like humans, AI systems can inherit biases from the data they are trained on. For instance, generative AI tools like ChatGPT can inadvertently favour certain cultural references over others, reflecting underlying biases in the data sets used for training. Moreover, AI-driven tools may not accurately assess students with diverse learning needs or those from varied linguistic and cultural backgrounds.
Educators, policymakers and technology companies must play a central role in eliminating and safeguarding against bias within these technologies. Many countries are beginning to define and implement regulations to ensure transparency and accountability in AI development and deployment but are being outpaced by the explosive proliferation of AI edtech tools.
Educators are pivotal in both advocating for unbiased AI and educating students on how to engage with technology, including AI. Some lessons from the way that resilient education systems helped students navigate a digital-first education during COVID-19-induced remote learning may also be instructive for the AI context.
For one, during periods of remote learning, teachers in the most resilient systems stayed highly engaged with students. They also continued to inform parents about their children’s progress, to ensure that parents stayed involved in their child’s learning. Resilient systems also seem to have invested into a solid foundation for student learning and wellbeing in schools, and increased peer-to-peer tutoring in school more than did all education systems on average.
The importance of the human aspect of education cannot be overstated, yet across the OECD on average, only 13% of students reported being asked daily, by someone from the school, how they were feeling. In Australia, 15% of students answered this in the affirmative (OECD, 2024b, 2024c). Educators must be available to support students who are feeling vulnerable and engage with them on an emotional level. If AI becomes more prevalent in the classroom, this human contact must be retained.
It is important to note that teachers are liable to have biases just as AI systems do. To overcome these biases, school systems should invest in cultural competency, implicit bias training, and equitable teaching practices. Additionally, implementing diverse and inclusive curricula, fostering open dialogues about bias and its impacts, and utilising data-driven approaches to monitor and address disparities in student treatment and achievement can create a more equitable educational environment, whether in a context of AI-empowered learning or analogue instruction.
Students need more support to recognise bias as well. The 2022 PISA reading literacy assessment found that less than half of 15-year-old students in OECD countries could distinguish facts from opinions when cues were implicit (OECD 2023c). Where AI tools are prevalent in the classroom, young people may come to trust their virtual interlocutors and be less likely to question any implicit or explicit biases. Teaching data literacy will help but so will basic literacy skills, too. Studies have shown that reading long texts is one of the best ways to impart a broad range of skills related to logical thinking, argumentation and contextual understanding – all important skills for helping students to navigate a digital information landscape.
Finally, technological advancements need to be complemented by strong parental support. Recent PISA data highlights a worrying decline in parental involvement in school activities between 2018 and 2022 across OECD countries (2023c). For instance, in 2022, only 29% of Australian students attended schools where at least half of the families proactively discussed their child’s progress with teachers, down from 40% in 2018.
This drop underscores the need to re-emphasise the importance of family engagement in education, especially as AI becomes more prevalent. In the context of AI education, where technology can sometimes create a sense of detachment, the human touch provided by family support is more essential than ever. Parents help bridge the gap between advanced learning tools and personalised attention, ensuring children receive a well-rounded education.
When parents actively engage with their child’s education, they provide emotional backing, reinforce the importance of learning, and help children navigate the digital learning environment. In 2022, 54% of students were in schools where teachers initiated discussions with parents, down from 59% in 2018, indicating a broader decline in proactive parental engagement (OECD, 2023c).
Data show that systems with more stable or improved parental involvement trends also exhibited better performance in mathematics (OECD, 2023c). This suggests that maintaining robust family support can mitigate the potential drawbacks of declining engagement.
The 2022 PISA results serve as a stark reminder of the challenges and opportunities presented by integrating AI into education. While AI holds the promise of personalised learning and improved educational outcomes, it also risks exacerbating existing inequities if not carefully managed. Ensuring equitable access to digital resources and providing robust support structures are essential steps towards harnessing AI’s potential for all students.
Additionally, addressing biases in AI tools and maintaining human elements in education – such as strong family involvement and emotional support from educators – are crucial in creating a balanced and inclusive learning environment. By prioritising these aspects, we can work towards an educational landscape where AI enhances rather than hinders equity, preparing all students to thrive in a digital world.
References
Bentley, S.V., & Naughtin, C. (2024, March 19). The ‘digital divide’ is already hurting people’s quality of life. Will AI make it better or worse? The Conversation. https://theconversation.com/the-digital-divide-is-already-hurting-peoples-quality-of-life-will-ai-make-it-better-or-worse-222987
OECD. (2023a). OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market. OECD Publishing. https://doi.org/10.1787/08785bba-en.
OECD. (2023b, December 5). PISA 2022 Results (Volume I and II) – Country Notes: Australia. https://www.oecd.org/en/publications/pisa-2022-results-volume-i-and-ii-country-notes_ed6fbcc5-en/australia_e9346d47-en.html
OECD. (2023c). PISA 2022 Results (Volume II): Learning During – and From – Disruption, PISA, OECD Publishing. https://doi.org/10.1787/a97db61c-en
In this article, Andreas Schleicher highlights the importance of family engagement in education. Thinking about your own school and students:
How often do you initiate contact with parents and families? Do you use these meetings to discuss student progress and learning? How often do you contact families to let them know about positive achievements and progress?
Our weekly newsletter with the latest articles, podcasts, videos and infographics from Teacher.
We carefully select offers of interest from our partners.
Facebook
Twitter
YouTube
SoundCloud
Apple Podcasts
Spotify
RSS feed
Linkedin
©2024 Australian Council for Educational Research — ACER, official partner of UNESCO