As Artificial Intelligence (AI) continues to evolve, its impact on education is sparking significant debate among policymakers and educators. The central question revolves around whether schools should adopt free, open-source AI tools or rely on proprietary systems that come with hefty price tags. This discussion not only highlights the digital divide but also raises concerns about accessibility and quality in educational resources.

Proprietary AI tools are often locked behind costly paywalls, potentially widening the gap between affluent institutions and those with limited resources. In contrast, free and open-source platforms promise broader access but may lack the necessary support and sustainability for long-term use. This dilemma echoes historical debates surrounding software adoption, where the free software movement advocated for user freedom while rejecting monopolistic practices.

To explore these issues, The Hindu hosted a live webinar titled "Choosing AI for education: Free or proprietary?" on September 6 at 5:00 p.m. The panellists included Bhanu Potta from ZingerLABS; Parth Shah from the Indian School of Public Policy; and Priyanka Kamath, Founder - 100 GIGA. It was moderated by M. Kalyanaraman from The Hindu's education vertical.

Human skills first

Priyanka Kamat emphasized the power of open-source in tackling challenges at scale. "We are building population-scale solutions with the young builders of our country, and we predominantly use open-source, because we can scale solutions better, especially when we are dealing with societal challenges that can be solved with technology," she said.

Looking ahead, she stressed that the real test for education lies not in competing with AI's technical skills but in nurturing distinctly human abilities. "In a world after AGI, where AI can probably do skill X or skill Y better than a human, what are we additionally bringing to the table for our students? EQ, creativity, critical problem-solving, on-field research -- these are the things we need to equip them with."

Ms. Kamat said that the push towards sovereignty in AI was non-negotiable: "We don't have a choice but to go forward with sovereign AI. India needs to go to a product and innovation mindset with sovereign AI versus just adapting from the West."

She illustrated this vision with an example from her own work: "We came up with this concept of skill currency, where for every module a student cracks, they earn reward points. Then they can use them to buy academic utilities, attend job fairs, or expand horizons with opportunities they otherwise wouldn't have."

AI evolution

Bhanu Potta broke down AI's layers of evolution. "There are three layers to AI: first, building the model; second, creating solutions on top of models; third, the agentic layer where AI can not only find but also act -- like not just finding cheap tickets but booking them on my behalf," he said. But he said, "For population-scale education, the real question is: how do I get this tool to the last public school in the remotest village, to the last child in my state?"

He cautioned against relying solely on market forces. "It's almost impossible to expect market players to build for everybody in the population. There is no other way but for institutions and State actors to come in and start thinking about sovereign models," he said. Instead of trying to overhaul the system all at once, Mr. Potta advocated for precision. "Rather than trying to boil the whole ocean of education, we should look at very surgical solutions: a dictation tool, a math practice tool -- cheap, scalable, effective." To him, technology design was only half the battle. "Apart from designing technology, we have to design adoption incentives -- otherwise equity cannot be achieved."

Bridging technology gaps

Parth Shah, drawing from his work at the Indian School of Public Policy, reflected on the analogy with past software debates. "AI is not like software. Software could be divided into categories -- word processor, data analysis -- and you had open-source and proprietary coexisting. AI does everything in one single bucket, so the analogy is different," he noted. Instead, he likened the current debate to India's long-standing struggle over medium of instruction. "Maybe the better analogy is language: the debate in India between English-medium and mother-tongue education is closer to the choice between proprietary AI and open-source AI."

Mr. Shah also highlighted a key difference in perspective between educators and employers. "As an educational institution, I care about learning, not just doing. Employers care only that you get the job done -- but I need to know whether my students are actually learning," he said. In his view, the real hurdle lies in the gap between possibility and practice. "There's always a gap between what technology makes possible and what human agency actually uses. That gap is often the real problem, not the technology itself."

Finally, he offered a note of caution against assuming only governments can deliver scale. "Every time a breakthrough technology comes, we think only governments can deliver it at scale. But if you look at history, it's often human motivation and incentives that decide adoption, not just availability."

As discussions unfold regarding effective implementation strategies for AI in classrooms, it is essential to focus not just on technology itself but also on enhancing teacher capabilities through targeted solutions. By empowering educators with robust tools tailored to their needs, we can create a more equitable learning environment that benefits all students.