Are our kids alright? Our use of artificial intelligence will tell

Straight from the pages of science fiction, AI and related technologies are a reality that we must face in the present day. But the history of emerging technologies calls upon us to ask: Is AI a harbinger of progress or doom?

“Yes.” This was the rather direct response Geoffrey Hinton, 2024 Nobel Laureate in Physics and dubbed “Godfather of AI”, gave when asked if artificial intelligence (AI) was smarter than humans in an interview from the American television program, 60 Minutes. 

AI grew straight from science fiction to a technological reality that all of us face today. It promises to revolutionize fields such as drug discovery, logistics, and education. As is the pattern for many emerging technologies in the history of science, its rising prevalence begs the question: Is it a harbinger of progress or doom? Is AI our friend or foe?

Whether we like it or not, there is no denying that it will be part of our society, at every level and where it matters. Since AI tools are designed to harness and manipulate data, the education sector stands out as a critical point of impact, especially now with the widespread use of Generative AI (Gen AI) and large language models (LLMs) like ChatGPT. With this, we have to take a step back and reflect on how AI technologies, specifically Gen AI, wield their influence over how our teachers manage the classroom and how our students learn. 

Talking about AI, it gets technical (a bit)

First, we must understand what AI is and what it isn’t. AI was born as a research tool for neuroscience. In order to comprehend how the human mind works, scientists modeled biological synapses with what are called “artificial neural networks” (ANNs). The seminal idea that inspired the whole field was seeded by Warren S. McCuloch and Walter Pitts in an academic paper, “A Logical Calculus of the Ideas Immanent in Nervous Activity.” At its heart, the paper forwards a hypothesis that the brain can be distilled into a series of computations and can be understood accordingly through computer systems.

The father of modern computing, Alan Turing, then attempted to address a fundamental question in computer science, “Can machines think?” by devising a thought experiment called the Turing test. This reconstructs the question into something more operational and manageable: “Could an intelligent machine and a human person be indistinguishable?” In other words, could machines imitate human intelligent behavior successfully enough to avoid detection or distinction?

With these organizing frameworks and the advent of modern computer science, we are now at the other end of a century’s worth of scientific progress, where computers can approximate the human learning process. This means the influence of several variables and, most critically, the correction of previous errors, are encoded into a machine’s decision-making processes. We now recognize it as “machine learning,” a feature of most modern AI platforms.

These are the operational principles behind the Gen AI tools we have at our disposal today, most prominently LLMs such as ChatGPT, which transforms textual inputs into organized informational outputs, the text-to-video generator platform Sora, and the larger AI ecosystem of Google Gemini.

Although AI holds unprecedented power as an information resource, it isn’t infallible. It is dependent on available information on the internet, as well as the level of training it receives on certain topics. It is also highly context dependent, as they are, indeed, language models, i.e., their outputs are strongly influenced by the textual input, as in the phrasing of a question. 

With this in mind, as with any sources of information, one should exercise caution in their use, and the veracity of sources should still be checked. This especially becomes a problem in education, especially with the ease and apparent reliability that Gen AI tools can provide.

Are our classrooms ready?

Recently, the Philippine Department of Education (DepEd) has inaugurated an initiative as an extension of its efforts to integrate technology into education—this time, involving AI. The Education Center for AI Research (E-CAIR) “will focus on revitalizing basic education by developing AI-driven tools that enhance teaching, learning, and school administration,” DepEd wrote in its press release.

In his keynote address, Sec. Sonny Angara ambitiously remarked, “Hindi po tayo makikipagsabayan para lang masabi na ‘tech savvy’ tayo. We are here to use AI as a tool for genuine, enduring reforms. This is the promise of the president’s Bagong Pilipinas.” Could the inclusion of AI tools become part of a solution to obstinate problems in local education?

The country, unfortunately, lags in terms of educational assessments compared to our ASEAN counterparts. The 2022 Program for International Student Assessment conducted by the Organization for Economic Co-operation and Development reported that Filipino students aged 15 scored below average in the fields of math, reading, and science. AI has already found its niche in multiple sectors of Philippine society, including advances in healthcare, biomedical science, and weather forecasting, arguably emerging as a bona fide technological revolution.

With the stark issues of education in the Philippines, however, ranging from curricular deficits to more logistical ones such as classroom shortages, it remains unclear how AI could truly bring upon educational reform. 

The question “Are we ready?” still echoes like a din with any new technology that arrives on our shores.

Due to the drastically enhanced access that AI tools could provide, it has the potential to accelerate accomplishing learning tasks by reducing the brunt of the workload that might be traditionally associated with the educational process, such as looking for references. It cuts the time that students might otherwise take to do tedious work and could potentially streamline education, such that other aspects or topics are attended to.

On the flipside, there is an obvious tradeoff, with the danger that quality is compromised for quantity and speed. One might also argue that it is precisely these more tedious aspects where learning—or rather, the development of important cognitive skills like critical thinking—takes place. Since AI tools enhance the convenience with which information can be accessed, cognitive faculties that are otherwise often utilized in traditional education might be diminished.

A recent study from the Massachusetts Institute of Technology shows exactly this effect, where students using LLMs to write essays were shown to accumulate cognitive debt, meaning a decline of their thinking capabilities and other neural processes. Using electroencephalography, which measures the brain’s signals, the study found that participants using their own brains to write the essay exhibited the strongest, most distributed neural networks, which are important for processing complex tasks. On the other hand, LLM users displayed the weakest connectivity, indicative of cognitive debt.

Later, participants were asked to return to using only their brains in producing the essays. Interestingly, LLM-to-Brain participants—those who transitioned from using an LLM to their own brains to write—showed the weakest brain connectivity, highlighting how one’s past reliance on Gen AI can influence future performance on cognitive tasks. 

“Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels,” the authors conclude. Although these results demonstrate marked negative effects, the paper, as of the time of writing this article, is still a preprint and has not yet undergone peer review. It is also worth noting that the paper contains a relatively small sample size.

Still, this is a troubling result that merits reflection on a more systemic level. What would AI-integrated learning look like in the Philippines, and how will this help bridge the ever-widening chasm in education? To answer this, we must look to the grassroots, the major stakeholders, and have expertise in the process: human students and teachers. 

Student attitudes toward AI

As AI boosts the accessibility of information, it is no surprise that students generally have positive attitudes toward it. A study conducted among 1,280 students in a Philippine state university revealed an overwhelmingly positive outlook on AI in enhancing the learning experience in terms of achieving learning outcomes. These were measured through metrics such as the ability of AI to enhance collaboration, pace their learning, personalize feedback, and keep up-to-date with field-specific knowledge. Top AI platforms that were used by participants in the survey were ChatGPT, Canva, Quillbot, and Grammarly.

Participants were also asked about ethical concerns surrounding AI, particularly on its effects on critical thinking, and a high average score was recorded, suggesting that students are aware of its negative effects. The survey also included concerns about fairness and transparency in education. It also found a positive correlation between a good outlook on AI and perceived usefulness, implying that students who found AI favorable were those who found it particularly useful. Surprisingly, an inverse trend was found among faculty, where no significant correlation was found in a similar metric, although faculty also generally had a positive attitude toward it. The authors of the study surmised that faculty attitudes toward AI might be influenced by other factors, such as professional development and ethical use, rather than utility alone. 

Perspectives of an educator

Ologist PH interviewed Jade Hizon, a professor of education at the Nueva Ecija University of Science and Technology. He teaches courses in the biological sciences and in education, and is passionate about integrating technology into the classroom. Drawing from his vast experience in the field of educational technology, which seeks to provide technical solutions to bridge educational gaps, Hizon has become a huge proponent of integrating AI tools into the curriculum. 

It is no secret nor is it a surprise that teachers are among the most overworked professionals, with so many roles to play, more than the already herculean task of providing the best quality of education for the Filipino youth. Teachers are overwhelmed with administrative tasks: from classroom management to navigating Philippine educational policies and their non-trivial gaps.

Hizon laments that an “exodus” is happening in the Philippine educational scene and believes solutions to this should start at the beginning of the process: the onboarding of new teachers. With this in mind, he and a collaborator started a project called SIBOL, meaning Smart and Supportive Induction for Beginning Teachers’ Onboarding and Learning. This AI-powered platform provides a professional dashboard for all concerns regarding the practice. This especially helps new teachers who have been in the practice for under three years to accomplish the demanding Induction Program for Beginning Teachers (IPBT).

“[The IPBT] has six modules that will introduce DepEd for beginning teachers […] It is really overwhelming for our beginning teachers. Hahanap ka talaga kung kanino ka magtatanong. Hindi ka makakuha ng konkretong sagot,” Hizon explains. “Imagine, [if beginning teachers] now have a personal coach that they can ask about the different infrastructures of DepEd.” SIBOL aims to prevent early career burnout in the workplace as it becomes a personal coach to talk to about the profession. This has the tandem benefit of replacing endless seminars as well as decreasing the workload of more senior teachers in becoming tireless professional resources for juniors.

Besides being a tool for professional guidance, Hizon has integrated AI in the way he approaches his teaching. He explains how it reduces the time needed to design assessments like quizzes and exams, as well as educational activities. He cautions against using it blindly, however, and recommends being more specific in the prompts to give the LLM when fashioning assessments.

While it is powerful, AI is not perfect. For instance, if insufficiently prompted, he observed that ChatGPT would create multiple-choice quizzes with mostly B’s and C’s as correct answers. This could easily be spotted by students and would take away learning opportunities and critical thinking skills. “Everything is faster, but that doesn’t mean that teachers aren’t supposed to read. We use AI as an extension of ourselves, as our extension of our cognitive asset. So hindi talaga dapat mag-rely 100%. So hindi dapat mawala ang touch natin,” Hizon adds.

With the burgeoning issues of AI and its potential effects on students’ cognitive skills, Hizon forwards a multipronged solution. He asks students to view AI as an extension of oneself, not a replacement, and to utilize it to create change for the better in their educational process. “Do not depend on it; listen to your teachers, even if you have ways to get it faster, do not forget the processes to do it the old way,” Hizon says.

Likewise, yardsticks must change in the advent of the Gen AI era, that is, students should be assessed differently. “Use AI with caution—ask students to present their output, kung ano ‘yung content. Paano nila nakukuha ‘yung desired output? Maybe we can ask our students to send the prompts, or ask AI to create an output and criticize the output of AI,” he says. Finally, he also brings up the critical role of parents and how AI capacity-building programs should involve guidance on how parents could use AI to facilitate their children’s education.

Friend or foe?

The bottom-line question remains: Is Gen AI friend or foe? 

AI is clearly a powerful platform for educational progress. When asked about the Philippines’ readiness for AI integration, Hizon answers, “Before readiness, we should be informed. Is everyone informed? What are we doing right now to inform people about AI? Hindi lang teachers and students, but parents as well.” He believes the Philippines needs to institute a systemic approach to AI like other nations. “To be honest, medyo kulang tayo sa initiative sa Pilipinas to have one national program for AI. In China, Europe, [AI is] integrated na sa curriculum.”

The Philippines needs to provide a policy-based agenda on how to responsibly integrate AI—especially generative AI—into our classrooms. Whether current initiatives such as E-CAIR would bloom into such a responsible and ethical AI ecosystem in Philippine education remains to be seen, as these are still in their infancy. Although large strides have been made in other fields such as healthcare, computer science, meteorology, biomedical sciences, and the larger fields of informatics, there is still much learning and work to be done toward integrating AI into the educational sector.

At its core, education is a human endeavor that molds and dictates our collective future. At the center of an AI revolution in the classroom should still be human ingenuity and intelligence, if we are to keep the values of critical thinking in the minds of our youth. Are we as a country asking the right questions about Gen AI and its influence in society? This will become one of the most important conundrums of the century. No better time to start thinking about them than now.