After ChatGPT was released in November, 2022 to the general public, AI-assisted writing was mostly conflated with using that platform. However, there are a wide array of tools available to students and LLMs are quickly permeating nearly every educational technology.
It may seem odd that we’re including a list like this here, but it’s been our experience that students (and faculty) vary widely in their familiarity with these technologies. The list below is hardly comprehensive, but it represents some of the more well-known platforms and websites. Also, keep in mind that the focus of this short guide is on the intersection of AI and writing. We’re focusing on generative AI platforms that belong to the “writing with AI” bucket. Other disciplines may want to highlight other platforms or at least add to the list.
Where to Find LLM Chatbots
Below is a list of the most commonly used Chatbots, as of Summer 2023. This list will be updated as new platforms emerge.
This is currently the most well-known LLM-based chatbot. It has a free and paid version. The free version is based on a less powerful version than the paid version (ChatGPT Plus). Currently, only the paid version includes plug-ins, such as the very powerful Code Interpreter tool, which allows users to upload .csv files for analysis, to get help with coding, and generate media through gifs and downloadable files.
Link: bing.com (currently only through Microsoft Edge browser)
After investing heavily in OpenAI, Microsoft integrated their GPT platform into nearly all of their most well-known products, including Bing. The result is Bing Chat, which has some functional similarities to ChatGPT but is missing other capabilities. Like Google Bard (below), it has browsing capabilities, an affordance that students can use to help with research.
Bing Chat (and Bard) have the ability to decipher and generate multimedia, not just text. Users can upload memes and other images to Bing Chat, for example, as part of a prompt.
A large bonus with Bing Chat is the sidebar experience Microsoft provides within the Microsoft Edge browser. When viewing a link or .pdf file, users can use the Bing Chat sidebar to interact with the text. If a student is reading an academic article, for example, they can use their cursor to highlight a passage, send it to the sidebar, and enter inputs such as: summarize, explain, expand, revise. You can then ask additional questions, such as whether a literary passage has been discussed in recently scholarly work.
This is Google’s version of ChatGPT. It has similar functionality to the free version. Like Bing Chat, it has multimedia capabilities.
Claude is an LLM from Anthropic. Currently, it has one of the largest “context windows” of any platform, which means that users can upload much larger amounts of text when prompt, as part of the context. Users can provide a file that contains the entirety of The Great Gatsby, for example.
Poe’s assortment of Chatbots
Poe is a platform developed by Quora that allows users to pick from a variety of chatbots, including limited access to the more advanced version of ChatGPT (that are otherwise only available through OpenAI’s paid tier). It also includes access to Anthropic’s Claude, Google’s PaLM (which undergirds Bard), and others.
Beyond Chatbots: How LLMs are transforming other web applications
While chatbot platforms are the most obvious place where students can leverage generative AI, LLMs are quickly permeating nearly all tools and educational technologies they will use throughout college and their careers.
Quillbot and Grammarly
Since this textbook is geared towards writing instruction, it’s important to become aware of how many tools that students have relied on have recently become suffused or amplified with LLM-based technology.
Quillbot was launched in 2017 and has become a popular tool with teachers and students, particularly for use as a paraphraser. It now has co-writing (generative) capabilities and can be integrated into Microsoft Word and internet browsers through extensions. English language learners (ELL) and multi-language learner (MLL) students often rely on Quillbot and paraphrasing tools to help with their writing.
Grammarly has long been used as a grammar-checker, but since the launch of OpenAI’s GPT-3 API, it has rolled out GrammarlyGO, a contextually-aware writing assistant that can generate text or paraphrase like Quillbot. Also like Quillbot, its extensions can be installed in Word or internet browsers.
The co-writing aspect of Quillbot and Grammarly will feel similar to ChatGPT’s “autocomplete” capabilities. On the other hand, these types writing-focused platforms offer more granular controls that may help students avoid unintentionally violating academic integrity at their institution. Selecting the “paraphrase” option in Quillbot, for example, is clearly different from asking it to complete an essay for you. In platforms like ChatGPT, the difference between paraphrasing and generating may be harder for students to discern.
Other AI Writing Products
The list of products that rely on generative AI is vast and growing. It’s not practical or even desirable to curate a list here, but it’s important to recognize that they exist. Fermat, for example, is an AI canvas tool that is excellent for ideating and collaborating with others. Sudowrite, on the other hand, is primarily used by creative writers.
Microsoft and Google Products
Both Microsoft and Google are implementing their LLM-based AI across their suite of products. Microsoft calls this AI capability Microsoft 365 Copilot.
Similarly, Google has begun including an AI-assisted writing option within Google docs, which works similar to prompting ChatGPT or Bard but will feel more limited. Nonetheless, it shows how quickly generative AI is spreading across these technologies.
Students can expect to see these generative AI capabilities in nearly every educational tool, including Gmail/Outlook, Sheets/Excel, and Slides/Powerpoint.
One way to sample Microsoft’s generative AI capabilities beyond Bing Chat is to try Microsoft’s Designer web app, which allows users to generate (via text-to-image capabilities) infographics based on simple text prompts.
Productivity and Research Tools
Many students use productivity tools, such as Notion, to take notes and practice time management. Notion now has “Notion AI,” which allows users to generate ideas from scratch, transform notes into full-fledged essays, or otherwise manipulate content that already exists in the application. Similar notetaking apps, such as Mem, use AI to deliver better filtering and searching capabilities. Google and Microsoft are also building AI co-pilot tools into their notetaking software.
Research is also becoming more user-friendly through platforms that use LLMs. Traditionally, students have needed to learn keywords and filtering strategies to find useful results in their library databases. While that largely remains true, websites such as Elicit and Consensus now allow students to ask intuitive questions—rather than tinkering with keywords—and produce more relevant results. Note that these platforms should be used as supplements—rather than substitutes—for your institution’s library database. And keyword searches will likely remain very effective. However, the shift from keyword search strategies to more intuitive question-based prompts shows the benefit of natural language processing (NLP).
Basic internet searches are also being transformed by LLMs. Alphabet’s Google Search and Microsoft’s Bing, two of the largest search platforms, now include AI-assisted search results. Also, websites such as Perplexity were early to integrate OpenAI’s technology into search and include AI search results with links to sources.
Coding, data analysis, and media generation for the masses
While much of the focus of this chapter is on text-based AI, the generative capabilities unleashed by current LLMs are opening new avenues for creative expression, even within first-year writing courses. Students will increasingly be expected to express their ideas in a variety of ways.
Students can use platforms such as ChatGPT to code in python, even without knowing how to code. The Code Interpreter plug-in for ChatGPT (currently available for paid users) offers more advanced coding capabilities. Users can upload excel files and ask it to perform data analysis. It can also generate .gifs. Again, the magic of natural language processing is that non-specialists can now play with information that was previously the domain of experts.
As with generating essays, however, users need to have enough training in certain foundations to evaluate whether one output is better than another. Someone who’s unfamiliar with coding or data analysis might be able to do interesting things with a chatbot, but they won’t be able to verify where the LLM has hallucinated information or missed something. Evaluating outputs is becoming an important skill.
Tips for Choosing a Platform
Generative AI can quickly feel overwhelming. The growth is fast-paced, and the space is not absent of tech charlatanry. How should you start?
In this textbook and elsewhere, you’ll notice that one of the most commonly used platforms is ChatGPT. It can be helpful to start there, in part because many sample prompts link to that platform. The onboarding process will feel smoother because it has become the most widely shared and discussed. As of June, 2023, users can begin sharing url links to their conversations for others to view and continue the initial prompts. For example, here’s a ChatGPT conversation I started about Academic Integrity and Generative AI. Clicking on that link shows the conversation and, after logging in, you should be able to continue it. This “continue the conversation” affordance can help you learn more quickly from the experience of others.
Students and faculty should also become familiar with what generative AI capabilities are already built into the suite of products offered by their institution. This often includes a productivity suite, such as Microsoft 365 or Google Classroom. As mentioned above, look for “co-pilot” options that are starting to appear in Microsoft Word and Google Docs. Both companies have announced these co-pilot options will appear across their entire productivity suite.
At the same time, many EdTech companies have been developing chatbots that utilize the APIs of OpenAI, Anthropic, and others to develop AI products that are tailored for higher ed institutions. A college may expect students to have training with these products. Khan Academy has worked with OpenAI, for example, to develop Khanmigo, a chatbot experience tailored for education. Your own institution may have partnered with a similar company.