OpenAI is releasing a teacher’s guide to ChatGPT, but some teachers are skeptical

OpenAI envisions educators using its AI-powered tools to create effective lesson plans and lessons for students. But some teachers are wary of the technology — and its potential to go wrong.

Today, OpenAI released a free online course designed to help K-12 educators learn how to bring ChatGPT, the company’s AI chat platform, into their classrooms. Created in collaboration with the non-profit organization Common Sense Media, with which OpenAI has an active partnership, the one-hour, nine-module program covers AI fundamentals and educational applications.

OpenAI says it has already released “several” case studies for schools, including the Agua Fria School District in Arizona, the San Bernardino School District in California, and the charter school system Challenger Schools. According to an internal company survey, 98% of participants said the program provided them with new ideas or strategies they could apply to their work.

“Schools across the country are facing new opportunities and challenges as AI reshapes education,” said Robbie Torney, senior director of AI programs at Common Sense Media, in a statement. “With this course, we are taking a proactive approach to supporting and educating teachers at the forefront and preparing for this change.”

But some teachers don’t see the system as useful – and think it can actually be misleading.

Lance Warwick, a sports teacher at the University of Illinois Urbana-Champaign, is concerned that resources such as OpenAI’s will make the use of AI common among educators who are unaware of the behavioral implications of technology. While the OpenAI course covers some of ChatGPT’s limitations, such as its inability to measure student performance, Warwick found the privacy and security modules “very limited” — and controversial.

“In the guidelines of the example [OpenAI gives]one tells you to enter grades and feedback from past assignments, and another tells you to create an assignment to teach the Mexican Revolution,” Warwick notes. “In the next safety module, it tells you not to enter student data, and then it talks about biases in AI production and problems with accuracy. I’m not sure if those fit the use cases. “

Sin á Tres Souhaits, a visual artist and teacher at the University of Arizona, says he has found AI tools useful in writing assignment guides and other supplemental learning materials. But he also says he’s concerned that OpenAI’s plan doesn’t address exactly how the company can use it to control the content that teachers already use its services.

“If teachers are doing courses and work for a program that gives a company the right to recreate and sell that data, that can be very disruptive,” Tres Souhaits told TechCrunch. “It’s not clear to me that OpenAI will use, package, or sell anything their models produce.”

In its ToS, OpenAI states that it does not sell user data, and that users of its services, including ChatGPT, own the results they generate “to the extent permitted by applicable law.” Without further assurances, however, Tres Souhaits is not sure that OpenAI won’t quietly change its policies in the future.

“For me, AI is like crypto,” said Tres Souhaits. “It’s new, so it offers a lot of possibilities – but it’s also so unregulated that I wonder how much I can trust any guarantee.”

Late last year, the United Nations Educational, Scientific, and Cultural Organization (UNESCO) forced governments to regulate the use of AI in education, including implementing age limits for users and regulations on data protection and user privacy. But little progress has been made in those areas since — and on AI policy in general.

Tres Souhaits also addresses the fact that the OpenAI program, marketed by OpenAI as a guide to “AI, generative AI, and ChatGPT,” does not mention any AI tools other than OpenAI itself. “It feels like this reinforces the idea that OpenAI is an AI company,” he said. “It’s a smart idea for OpenAI as a business. But we already have a problem with these tech-opolies – companies that have a huge impact because, as technology is developed, they put themselves at the center of innovation and make themselves the same thing. “

Josh Prieur, a classroom teacher turned product director at educational game company Prodigy Education, was more excited about OpenAI’s reach for educators. Prieur argues that there are “obvious benefits” for teachers if school systems use AI in a “thoughtful” and “responsible” way, and believes the OpenAI initiative is transparent about the risks.

“There are still concerns from educators about using AI to copy content and undermine the learning experience, and the dangers of overreliance on AI,” Preiur said. “But education is often the key to overcoming fears about the adoption of new technologies in schools, while also ensuring that the right safeguards are in place to ensure that students are protected and teachers remain in full control.”

OpenAI is aggressively pursuing the education market, which it sees as an important area for growth.

In September, OpenAI hired former Coursera CEO Leah Belsky as its first GM of education, and charged her with bringing OpenAI products to more schools. Also in the spring, the company launched ChatGPT Edu, a version of ChatGPT designed for universities.

According to Allied Market Research, the AI ​​in education market could be worth $88.2 billion over the next decade. But growth began slowly, in large part because of skeptical teachers.

In a survey conducted by the Pew Research Center this year, a quarter of public K-12 teachers said that using AI tools in education is doing more harm than good. A separate survey conducted by the Rand Corporation and the Center on Reinventing Public Education found that only 18% of K-12 educators use AI in their classrooms.

Education leaders have been similarly hesitant to experiment with AI themselves, or to introduce the technology to the educators they look after. According to education firm EAB, several district superintendents view addressing AI as a “highly urgent” need this year — especially because of pressing issues like staff shortages and chronic absenteeism.

Mixed research on the educational impact of AI has not helped convince the naysayers. Researchers at the University of Pennsylvania found that Turkish high school students with access to ChatGPT performed worse on a math test than students without access. In a separate study, researchers found that German students who used ChatGPT were able to find research materials more easily, but they tended to synthesize those materials with less skill than their peers who did not use ChatGPT.

As OpenAI writes in its guide, ChatGPT is not a substitute for student interaction. Some teachers and schools may never be convinced that it can replace any step in the teaching process.