AI is coming for students

It is officially back-to-school season, and this year’s school halls are welcoming students, educators, and lots of AI -- or at least that's what major AI companies seem to be aiming for. The most recent wave of AI releases shows a shift in how companies are positioning chatbots as more than just answer machines, but tools that can even encourage learning. 

This week, Anthropic launched a new Learning Mode available to everyone in its Claude.ai chatbot and Claude Code, the coding assistant that helps developers within their terminals or IDEs. In Claude.ai, it helps people arrive at the answer they are seeking through guided questions, mimicking the Socratic method. In Claude Code, it leaves fill-in-the-blank spaces for a collaborative coding process. An additional Explanatory mode makes Claude explain what it did and why, which Anthropic likened to a senior developer talking through their process. 

OpenAI released its own Study Mode two weeks ago, which similarly pivoted from just providing answers to working with the user to arrive at the conclusion. I tried the feature and noticed a big difference between the responses in standard ChatGPT and Study Mode, especially when I fed each an AP trigonometry problem. ChatGPT immediately gave me the answer, while Study Mode made me work for it, and, when I wanted to give up, even encouraged me to keep working through it. However, whether this experience translates into actual learning is an entirely different story. As ZDNET’s Tiernan Ray, who also went hands-on with the feature, noted, “with all language models, you only get out of it what you put into it…In other words, education in Study Mode results from the student's effort rather than the teacher's brilliance.” 

That isn’t stopping companies from marketing their AI tools to students. As part of its back-to-school release, the Speechify text-to-speech app launched a competitor to NotebookLM’s AI-powered podcast tool. I tested it and was pretty impressed. Last week, Google made its AI Pro plan, which includes all of the best of Google's AI suite of tools, free to college students (it's normally $20 per month). 

The applications of AI in the education sector are so nascent that it isn't possible to know yet if the risks -- including hallucinations, misinformation, or AI-enabled cheating -- outweigh the benefits, which include having access to an endless amount of information, like a tutor, available at all times. Even with the upsides, educators and students alike are concerned AI will erode critical thinking skills, which research shows is possible.

AI companies are continuously improving their tools, which presumably will mean a better experience for all users. Anthropic and Google just updated the Claude and Gemini chatbots, respectively, with a feature similar to ChatGPT’s Memory, meaning the bots can reference your past chats for more personalized responses. Also notable: Google just gave the Gemini Live app Google Calendar, Keep, Maps, and Tasks support, xAI gave users 'limited' free access to Grok 4, and Microsoft's new Copilot 3D turns a 2D image into a 3D model. 

Speaking of upgrades, despite all the initial hype and anticipation, OpenAI’s launch of GPT-5 has gotten off to a rocky start. Upon release, GPT-5 replaced all legacy models in the model picker, aiming to be a one-stop shop for most queries. But when paid subscribers were disappointed by the disruption to their workflows, OpenAI had to bring older models back.

User feedback has been mixed. GPT-5’s coding capabilities impressed some but underwhelmed others, including ZDNET’s David Gewirtz, though he did say GPT-5 Pro delivered sharp, actionable code analysis. OpenAI’s claims that GPT-5 is more accurate than any previous reasoning model and has fewer hallucinations have aligned with lots of user feedback. Still, the release has some questioning the possibility of superintelligence.