The Illusion and Appeal of LLM Reasoning

Words like reasoning, thinking, and writing are the working tools of the legal profession. But with the rise of large language models, like OpenAI’s GPT, Anthropic’s Claude, and Google’s Gemini, these words are now used in a different way. If we don’t confront their false familiarity, we risk misunderstanding the capabilities of these tools and misplacing our trust in them.

Using Generative AI to Discover What You Don’t Want

Writers often discuss the zero draft—a rough document where they begin shaping their ideas. Anne Lamott calls it the “shitty first draft.” But what if getting to that point feels impossible? Before reaching the zero draft, writers can try something new: the negative draft.

Introduction to the New(ish) Technology Competence

Technology competence is not new, but in the age of generative artificial intelligence (GenAI), it applies in new, more complex ways. Under ABA Model Rule 1.1 and its Comment 8, competent representation includes understanding the benefits and risks associated with relevant technology. When it issued ABA Formal Opinion 512, the ABA reaffirmed that the duty of technology competence applies to GenAI, along with all other ethical duties, including:

Since generative AI can produce polished text in seconds, it’s tempting to ask: Why write at all? If the end result, like a report, an email, a memo, or a presentation, looks the same whether written by a human or GenAI, why not just let the technology do the work? Because writing isn’t just about producing text. Writing is thinking.

It may seem efficient to use generative AI (GenAI) tools to write content for you. You’re busy. Maybe you’re not deeply invested in the final product. Maybe you just want to be done. However, GenAI often produces text that is bland, abstract, repetitive, obvious, and just awkward—especially compared to a human writer who knows the topic well.

Something feels off about your new robot co-worker—besides the fact that your co-worker is a robot. This robot produces grammatically correct text at lightning speed. The writing seems natural, not robotic. It’s impressive, but is this text good and should you adopt it as your own?