Teaching in the age of AI means getting creative

Alarm bells seemed to be ringing in teacher halls across America late last year with the debut of the first school chat – An easy-to-use AI chatbot capable of producing dialogue-like responses, including writing and longer articles. Some writers and teachers have gone so far as to predict Death of student papers. However, not everyone was convinced it was time to panic. Many naysayers pointed to the bot Unreliable resultsAnd factual inaccuracy And dull toneinsisted that technology will not replace fact writing.

Indeed, ChatGPT and similar AI systems are being used in fields beyond education, but the classroom seems to be where concerns about bot abuse — and ideas to adapt alongside evolving technology — begin first. The realities of ChatGPT force professors to take a hard look at current teaching methods and what they actually offer students. Existing types of assessment, including core articles that ChatGPT can emulate, may become obsolete. But rather than calling AI a gimmick or a threat, some educators say this chatbot could end up resetting the way they teach, what they teach and why they teach it.

At Santa Clara University this month, 32 students started a course called “Artificial Intelligence and Ethics” in which the usual assessment method — writing — is no longer in use. The course is taught by Brian GreenInstead of essays, he will run one-on-one sessions with each student for ten-minute talks. He said that evaluating this takes no more time than grading an article.

“In that context, you really remove any possibility of script generation software. And when you talk to them, it really comes down to whether or not they understand the material.

But such an approach may not be realistic in all educational contexts, particularly in schools where resources are scarce and teacher-student ratios are worse.

At some universities, the response to this technology has simply been to restrict access. Earlier this month, the New York City Department of Education announced that ChatGPT would be so Blocked on networks and devices throughout its public schools. “While the tool may be able to provide quick and easy answers to questions, it does not build critical thinking and problem-solving skills, which are essential for academic and lifelong success,” a department spokesperson said in a statement.

And the nation’s largest public school system isn’t alone: ​​Teachers at various levels around the world have it They aired their fearsand other counties in the United States, such as Seattle public school systemthey also restricted technology.

But such a ban is not a solution. Anyone with access to a smartphone — like 95 percent of Americans ages 13 to 17, According to a Pew Research Center survey last spring These restrictions can be easily bypassed without the need for a school computer or campus Wi-Fi.

Some teachers told FiveThirtyEight that they see ChatGPT bans as misleading responses that misunderstand what the tool can and cannot do.

“ChatGPT may have better syntax than humans, but it’s shallow in terms of research and critical thinking,” he said. Lauren GoodladProfessor of English and Comparative Literature at Rutgers University and Chair of the Critical Artificial Intelligence Initiative. She said she understands the concern about the tool but that – at least on a college level – the type and caliber of written assignments that ChatGPT can offer do not replace critical thinking and human creativity. “These are statistical models,” she said. “And so they prefer probability, since they are data-trained, and the only reason they do as well as they do is because they look for potential responses to the prompt.”

These point to limitations that stifle chatbots’ originality, such as how statistical models favor the use of more common words at the expense of rarer words that human authors might use. Gödel also pointed out that the tool is not always accurate at this time. For example, ChatGPT is vulnerable to “hallucinationsor provide false sources and quotes.

It’s those kinds of tags that might help teachers currently not only catch students trying to pass off ChatGPT-generated text as their own, but also put in place measures that encourage students to do the work themselves. Some suggestions The ones she and her colleagues explain include asking students to reference class discussions in their work, attaching a reflective video or blurb on why they chose the writing points they did, and asking that specific oratorical skills appear in the essay.

But it’s more important, Goodlad said, that schools improve by changing what they emphasize in their curricula, suggesting that teachers turn instead to teaching methods and written assessments that emphasize critical thinking. Otherwise, these methods can quickly become outdated.

“The whole space has basically become an arms race,” Green said, adding that anti-cheat technology is still in constant competition with technology to circumvent it, as has been the case for years with plagiarism detectors like TurnItIn. The dynamic with ChatGPT will probably follow the same pattern. For example, earlier this month, Princeton University student Edward Tian revealed that he had developer program To detect work written in ChatGPT. While the news has received some praise, many see it as just a temporary measure.

“These tools are only going to get more advanced,” he said. Hood LipsonProfessor of Mechanical Engineering and Data Science at Columbia University. “This is no different from the beginning of the Internet or Wikipedia. And it was a mistake to prevent students from using Wikipedia or Google search, right?” He said the question is not whether the technology should be banned but how to evolve alongside it.

Lipson tries to incorporate ChatGPT and similar technologies into his education. For example, in his introductory robotics course this semester, he will ask his students to use DALL-E — an image generation program developed by OpenAI, ChatGPT’s parent company, and powered by similar technology — to help think about their initial drawings for a project. robots that they will work on throughout the semester. “With just a few keywords, it takes the machine about 25 seconds to generate maybe 25 designs or concepts — something that would have taken students a week to produce,” he said.

Rather than blocking, then, the future of teaching may be a mixture of new approaches using tools like ChatGPT and older approaches – pen and paper tests, like some Australian universities are back Help regulate students’ dependence on technology.

And many educators, regardless of their current ChatGPT curriculum, continue to build on optimism that such technology will eventually push us to get to the heart and soul of what education means, focusing on deeper understanding rather than simply developing a skill.

“We know that calculators exist,” Green said. “But we still study math.”