Recently, I gave a conference to a group of teachers from kindergarten to the 12th year and public school administrators in New York. The subject was artificial intelligence and the way schools should adapt to prepare students for a future filled with all kinds of capable AI tools.
But it turned out that my audience was concerned with a single AI tool: Chatgpt, the Buzzy chatbot developed by Openai which is able to write convincing tests, to solve problems of science and mathematics and to produce of the work computer code.
Chatgpt is new – he was released at the end of November – but he has already sent many educators in panic. Students use it to write their homework, passing tests and sets of problems generated by AI. Teachers and school administrators rushed to catch the students using the tool to cheat, and they are worried that the ravage cat could carry out their course plans. (Some publications have declaredPerhaps a little prematurely, that Chatgpt has completely killed homework.)
Cheating is immediate and practical fear, as well as the propensity of the bot to spit erroneous or misleading responses. But there are also existential concerns. A high school teacher told me that he had used Chatgpt to assess some of his students’ documents and that the application had provided more detailed and useful comments on them, in a tiny fraction of time.
“Am I even necessary now?” He asked me, only half a joke.
Some schools responded to Chatgpt by repressing. New York public schools, for example, recently Blocked Chatppt Access On computers and school networks, citing “concerns about negative impacts on student learning and concerns about the security and accuracy of the content”. Schools from other cities, including Seattle, also have limited access. (Tim Robinson, spokesperson for Seattle public schools, told me that Chatgpt had been blocked on school aircraft in December, “with five other cheating tools.”)
It is easy to understand why educators feel threatened. Chatgpt is an incredibly capable tool that landed in the middle of them without warning, and it works quite well on a wide variety of tasks and academic subjects. There are legitimate questions about the ethics of writing generated by AI and concerns as to whether the answers that the cat cat is accurate. (Often, this is not the case.) And I am in favor of teachers who feel that they have enough to worry, without adding homework generated by the AI to the mixture.