AI tutors could provide students with two or three times more benefits than traditional classroom learning, according to studies cited by Russell. He said there may be “reasonable fears” of job loss among teachers and their unions, but he is focused on the potential added value from AI tutors rather than robots replacing them.
Stuart Russell, a professor of computer science at the University of California, Berkeley, and an AI expert, predicts that AI-powered chatbot tutors could revolutionize traditional education with their one-on-one training benefits.
Last year, ChatGPT created a buzz among students who embraced its ability to replicate human conversation. As technology advances, these personalized tutors may become increasingly commonplace.
AI has the potential to deliver high-quality education to all children in the world by the end of this decade, according to Russell, a proponent of its use. “That’s potentially transformative,” he said.
Russell, who was at the United Nations’ AI for Good Global Summit in Geneva, suggested that personalized chatbots could potentially provide an educational solution for high school-level material, accessible from mobile phones or computers.
OpenAI is currently testing a virtual tutor program powered by GPT-4, in collaboration with an education nonprofit, which it says functions as both a tutor for students and a classroom assistant for teachers.
AI tutors could provide students with two or three times more benefits than traditional classroom learning, according to studies cited by Russell. He said there may be “reasonable fears” of job loss among teachers and their unions, but he is focused on the potential added value from AI tutors rather than robots replacing them.
A spokesperson for the National Education Association said in a resolution at its annual meeting that the use of artificial intelligence in the classroom should support students’ and educators’ needs, be equitable and accessible, and not contain any bias.
Russell argued for the potential use of AI in education, though acknowledging risks such as indoctrination. He suggested that human involvement could morph into roles such as “playground monitor” or educator on topics like civics and morality.
A study released in March, which hasn’t been peer-reviewed, found that ChatGPT was able to answer questions in MIT undergraduate courses with 100 percent accuracy. But three students examining the methodology noted “glaring problems” which could have been used for cheating. The authors of the study included 15 experts, including MIT professors.
Russell has warned of the dangers posed by AI, urging caution despite its growing use since ChatGPT’s release. Thousands of experts, leaders, and others signed an open letter calling for a pause in AI research until safety protocols are established. Russell said advanced AI design must be accompanied by shared safety protocols to ensure it helps humanity flourish and not cause “terrible harm.”
