What are Generative AIs and why are they suddenly changing?

Generative Artificial Intelligence (GenAI) technologies have garnered significant media attention for their ability to produce high-quality text and images in response to written prompts. Since the public launch of several large language models in 2022, GenAI has raised considerable concerns about academic integrity and the future of writing as a learning tool. While some of the initial hype and panic may be abating, navigating the landscape of writing in the era of GenAI will require new ways of thinking about authorship, intellectual property, and how our writing technologies shape our writing lives.

What is GenAI? Is it like we see in the movies?

Generative artificial intelligence tools are based on foundational technologies called large language models (LLMs). While Google Gemini, Anthropic’s Claude, Microsoft Copilot, and OpenAI’s ChatGPT have garnered the most initial media attention, several large tech companies (Meta, Apple, and others) have LLMs that are increasingly integrated into the tech tools we use every day. Indeed, most word processing, document reading, and search tools now offer opportunities to use AI for search, brainstorming, writing, revision, and reinvention of written text. Agentic AI tools go even further, integrating with other sites and software on a device to independently complete complicated multi-part tasks based on a single prompt. Many companies are developing specialized AI models for customer service, marketing communication, law, medicine, and other fields. Despite the proliferation of many tools and technologies, a smaller number of proprietary large language models power their operation.

Like a parrot that can mimic human speech in response to cues, LLM outputs are probabilistic and limited by the characteristics of their training sets. Emily Bender and colleagues described LLMs as “Stochastic Parrots” in their 2021 article on the dangers of these technologies, and the metaphor remains valuable. The models are stochastic, meaning they rely on statistical prediction, along with all the affordances and challenges that statistical estimation entails. Unlike human text production, which is based on meaning and memory, LLMs convert the ‘natural language’ of a written input into code, which is then assessed according to parameters, leading eventually to an output that is either written text or a generated image. LLMs do not comprehend meaning or understand words, sentences, or paragraphs; rather, they generate output through massive statistical calculations based on pattern recognition. In subsequent work, Bender refers to these tools as synthetic text-extruding machines.

This “meaning gap” produces hallucinations: counterfactual claims in which a Gen AI tool produces output that is inaccurate or unrelated to reality. Because these models gather small pieces of relevant information but don’t understand meaningful connections among them, the statistical patterns may not reflect reality. For example, a hallucinated citation may identify a real author and attribute their work to an actual journal, but details such as the title, volume, and page number could refer to something that doesn’t exist.

To improve output quality and reduce errors, most AI developers are leaning heavily into the LARGE of large language models. While the initial models used millions of statistical parameters to identify relationships among words, the largest models now use trillions of such relationships. This expansion is driven by ongoing increases in computing power and rising demand for power, water, and other resources.

Where did GenAI come from? Does it matter?

Generative AI has evolved from two areas of computational research: natural language processing and machine learning. Natural language processing research aims to enable more effective human/computer interaction by helping technologies interpret how ordinary humans use speech and writing. Machine learning, by contrast, aims to create technologies that operate more or less independently of human intervention by developing systems that can test and improve themselves without direct programmer involvement. These roots are clear in the development of Microsoft Copilot, conceived as a tool that enables programmers to use an AI code-writing assistant to create new tools and technologies.

A recent breakthrough in these technologies emerged with the development of transformer architectures (in fact, the GPT in ChatGPT stands for Generative Pretrained Transformer). Transformers “learn" to create generalizations based on a training dataset by encoding and decoding samples. The first generation of transformers was trained to recognize and sort items into classes (so-called discriminative models). In contrast, generative transformers use the conventions learned from examples to generate probable examples. ChatGPT is pre-trained on massive statistical parameters (175 billion) to examine a gigantic quantity of text. Based on its analysis of prior human-written discourse, it generates text that closely resembles the output of a real writer. Most GenAI models operate on similar architectures but vary in the number of parameters and the content of their training sets.

While the technical details of GenAI tools move quickly into the highly specialized domains of computer programming, it’s valuable for faculty to understand that GenAI technologies are not entirely new, nor are they entirely different from the computational work that brought about ubiquitous low-cost computing, the internet, and aggressive attention to algorithms. Challenging both the utopian visions of AI marketing teams and the nightmare scenarios imagined by critics, researchers Arvin Narayanah and Sayah Kapoor suggest it is wisest to think of AI as a “normal technology.” In most cases, GenAI tools promise increased productivity and efficiency through automation, intending to augment or replace human labor. 

Is all this technical detail important for instructors?

Some technical understanding of how LLMs work is vital for both users of GenAI and anyone who may encounter AI-generated content online for several reasons:

• Materials provided by companies about their tools may assume a certain level of technical knowledge, and understanding the basics of training, prompting, and data collection can make potential users more informed and effective.
• Users who understand how Large Language Models work can more effectively distinguish between tools and products.

While Grammarly, Khanmigo, and Perplexity offer their services for a fee, each uses ChatGPT as its primary engine for generating results. The University of Minnesota provides enterprise access to two of the largest LLMs, Google Gemini and Microsoft Copilot, and each offers versions with different capabilities and resource demands.

• Resisting the temptation to anthropomorphize Generative AI can help us recognize that these tools are powerful, yet limited.

Those marketing GenAI tools promote the sense that AI is an interlocutor or companion, and those anthropomorphic representations can generate a great deal of misplaced trust. Novices are easily fooled by the language choices and speed of GenAI output, and often overvalue the quality of AI output. Indeed, evidence suggests that strong dependence on AI models contributes to deskilling and neverskilling in educational contexts.

• Understanding how these tools operate will be crucial for determining when AI use may benefit students and when it may hinder learning. 

Few instructors bemoan the decline of students’ spelling competency in the wake of spellcheck. At the same time, learning researchers are very concerned about the propensity of these tools to shortcut the critical thinking and skill acquisition that are at the core of higher education. Students require practice in information-seeking, reasoning, and writing to develop transferable skills for work and life.