ChatGPT is not the only AI writing tool

ChatGPT has been receiving a lot of publicity since its public beta launch in November 2022. But using AI (specifically ‘large language models’ [LLMs]) for some writing tasks has been around for a long time.

At least 10—possibly 20—years ago, I heard of machine translation (MT) tools when I was working as a technical writer in the computer software industry. For those having to translate user manuals into multiple languages (as required by the EU etc.), these tools were essential for dealing with the repetitious stuff (translate once, re-use many times). Sure, they had their issues, but over time, and as the ‘machine’ learned more, they got much better. Ideally, in-country human translators were then used for checking and cleaning up, and feeding changes back to the ‘machine’ for it to learn better for next time. As each new version of the software was released, only the new information needed to be translated because the MT tool stored the previous changes and had ‘learnt’ the patterns, thus saving companies millions of dollars in translation costs (with software releases generally being about every year, you can just imagine how much it would cost in time and money for human translators to do EVERY translation into 20 or more languages for EVERY release).

Just to be clear, I’ve never used any of these tools or was ever involved in translation projects, so what little I’ve written here is based on what I learned at tech writing conferences etc.

Because the discussions around ChatGPT have interested me, I’ve taken note of some articles, especially related to how to use ChatGTP as a tool amongst the suite of tools I use for editing. Others have spent plenty of time writing about ‘the sky is falling’ scenarios in schools and universities, and some have written ‘let’s all just take a deep breath and see how this can work for us’ articles to counter them.

One thing I’ve discovered in the past few days is that there are PLENTY of AI tools around, not just ChatGPT, that summarise text and do other other writing and editing tasks. And some of these tools have been around for several years—they just haven’t received the publicity and hype that ChatGPT has. (Update 31 January 2023: This website attempts to list many that are publicly available: https://www.futurepedia.io/)

Text summaries provided by AI tools could be used for creating an executive summary of a report, unpacking dense writing, summarising a ‘too long, didn’t read [TLDR] article into the main points for a busy person to read, etc. Some of these tools (plus others) are mentioned and linked to in these articles if you’d like to investigate them further:

• A Brief Introduction to Text Summarization (Anne-Maj van der Meer; 19 Dec 2022): https://www.taus.net/resources/blog/a-brief-introduction-to-text-summarization
• AI and Scholarly Publishing: A View from Three Experts (Anita de Waard; 18 January 2023): https://scholarlykitchen.sspnet.org/2023/01/18/guest-post-ai-and-scholarly-publishing-a-view-from-three-experts/

And an academic article on LLMs: Dissociating language and thought in large language models: a cognitive perspective
(Kyle Mahowald, Anna A. Ivanova, Idan A. Blank, Nancy Kanwisher, Joshua B. Tenenbaum, Evelina Fedorenko; 16 January 2023): https://arxiv.org/abs/2301.06627

[Links last checked January 2023]