.png)
Hey, curious minds! Let's dive into the captivating realm of AI tutors and the skepticism surrounding them, with a glance back at IBM's 'Watson' era. Strap in for a rollercoaster of insights into the possibilities and challenges that lie ahead!
Before ChatGPT became the talk of the town, there was a team of computer scientists at IBM, led by Satya Nitta, embarking on an ambitious mission – crafting a revolutionary personal digital tutor using the power of artificial intelligence. This was a time when AI wasn't making as many headlines, and IBM's Watson was stealing the show by outsmarting humans on Jeopardy in 2011.
Nitta and his team, fueled by optimism, believed that Watson could be the driving force behind a generalized tutor. However, he acknowledged the Herculean task at hand, stating, "I remember telling IBM top brass that this is going to be a 25-year journey."
After about five years of dedicated efforts, including contributing to learning products like a chatbot assistant in a Pearson online psychology courseware system, Nitta reached a verdict. Despite the groundbreaking capabilities of generative AI, he concluded that it's not yet equipped to fulfill the role of a generalized personal tutor. According to Nitta, "We'll have flying cars before we will have AI tutors."
Undeterred by the challenges, Nitta co-founded Merlyn Mind, a new AI company focusing on building different AI-powered tools for educators.
While some experts like Nitta are skeptical, the race to create AI tutors is still on. Sal Khan, the leader of Khan Academy, even envisions AI as the catalyst for the most significant positive transformation in education. Khan's Khanmigo, an AI tutor in pilot phase, aims to provide every student with a personalized learning experience.
However, the road to AI tutoring is not without its bumps. The "hallucination" problem, where chatbots fabricate details when faced with unknown queries, poses a significant hurdle. Efforts to address this include employing separate AI chatbots to verify information, though achieving the desired accuracy is an ongoing challenge.
In the midst of these developments, Michael Feldstein, an ed-tech consultant, offers a pragmatic perspective. He acknowledges ChatGPT's capabilities but emphasizes the importance of not overselling the limits of AI tools. Drawing from personal experience, he shares how ChatGPT has served as an educational resource, tutoring him on a serious medical condition.
Feldstein believes that while these tools may not be all-knowing, they can serve as an "always-on helper" for users seeking information and learning. He advocates for viewing AI tools as valuable aids rather than flawless entities.
Satya Nitta leans towards AI serving as an assistant to experts, automating tedious tasks rather than replacing human tutors. Merlyn Mind's focus on building AI tools to enhance educators' capabilities aligns with this vision. Nitta emphasizes the importance of tailored training for AI models in specific education domains.
As the journey towards effective AI tutoring continues, the question arises – is tutoring the right goal, or can generative AI serve a different purpose in education? Whether it's as an "always-on helper" or an assistant to experts, the evolving landscape of AI in education promises exciting possibilities.
What are your thoughts on AI tutors? Are they the future of education, or do we need to redefine their role in the learning process?
Comments