When we first heard about the new generative artificial intelligence (AI) program that was taking over the internet, we thought it was too good to be true. Artificial intelligence has been growing for years, and you might already use AI in several facets of your classroom. But new concerns for educators are developing around the increasing popularity and accessibility of AI bots and tools.
ChatGPT, a new AI writing tool, was released in November 2022 as a prototype and has been buzzing around social media for weeks. ChatGPT is designed to create text responses in a manner that is realistic to human speech, known as a generative pre-trained transformer (GPT). ChatGPT can solve math problems, write poems and essays, generate computer code, and more. With chatbots being more accessible, there are many potential benefits:
Though we haven’t seen a chatbot of this caliber accessible to the public before, it seems that AI chatbots are not yet at a safe and reliable level. Many limitations have been found in ChatGPT’s services, some of which are even acknowledged by OpenAI right on their landing page:
Like millions of users worldwide, we had to test the website for ourselves, and our team was impressed by the general quality of responses it produced. The chat responses had natural syntax, followed genre guidelines, and integrated prompt instructions well. The system also modified responses if prompted.
But the system was indeed too good to be true: we were disappointed in the vagueness, inaccuracy, and questionable sources the platform created. When we asked the chat to create an essay outline, the generated outline lacked a compelling depth. We also asked the chat to answer some division problems, to which some were incorrect. The site experienced several network errors while using the chat.
AI technology that can generate almost any response you ask must be taken seriously, especially in education. ChatGPT has many educators concerned about cheating, plagiarism, and the ability of students to be critical and creative thinkers. One article from The Atlantic called chatbots “the end of high-school English”. Several school districts, including New York City Department of Education, have banned the site from school computers and networks, while many educators believe these districts acted too quickly.
But it seems the tides are turning as more educators have no choice but to embrace AI. Some educators have already taken advantage of the site by allowing students to brainstorm and generate essay outlines. There’s potential for chatbots to be a great place for brainstorming, learning about genres, and practicing outlines.
Two months after its release to the public, OpenAI released a tool to help educators detect whether their system generated a piece of writing. When an educator inputs a specific text, AI Text Classifier responds to inquiries with an evaluation including “very unlikely, unlikely, unclear if it is, possibly, or likely.” The company admits that, just like ChatGPT, their AI Text Classifier isn’t perfect. Jan Leike, who is leading efforts to make their AI systems safer, says their classifier “will be wrong sometimes.”
While ChatGPT has become wildly popular, the race to create more safe and usable AI chatbots has officially begun. On February 6th, 2023, Google announced that they are soon launching a “rival” chatbot called Bard. The announcement was met with criticism as the bot provided incorrect information in the announcement promotion. The same week, Microsoft announced that ChatGPT would soon be incorporated into Bing’s search engine responses. The BBC called this move the “biggest threat” to Google’s domination of internet searching. Baidu, the tech giant in China, will also be releasing an AI chatbot in March 2023. But we fear that the intense rush for big tech companies to release these chatbots will increase the likelihood of imperfect programming that leaves millions with access to dangerous tools.
Now we’re left asking: how do we adapt to AI moving into our classrooms? Our students might soon be equipped with a powerful tool that could change the direction of their education for better or worse. Our team considers it part of our job as educators, especially as AI becomes more usable, to shape independent students and critical thinkers. Critical thinking teaches students how to make decisions, improve relationships, be creative, build independence, and problem-solve—all skills required for personal and academic success in the modern world. As AI improves, more adjustments will be needed in our administration and our classrooms, but the time to start educating students on informed perspectives around AI has been pushed into our laps.
While AI chatbots become more widely accessible, Mobile Ed has a few practical recommendations to help you promote student success in your classroom in the age of AI:
Sign up for the Mobile Ed blog for more tips for educators.