top of page
Writer's pictureKim Melton

AI-yai-yai: Using GenAI to Create Technical Content

Updated: Dec 13, 2023

A black and white image of a person with a human brain overlaying their head and a sphere of connected points around it







ChatGPT - are you sick of hearing about it yet? AI in general isn’t just getting a lot of press, it is dominating headlines across industries. That’s understandable. AI is creating rapid changes to markets throughout the world, from the medical field to graphic design. This has everyone sitting in a weird place between is my job at risk? and how can I use this to make my job easier?


While it’s clear that we’re several years out from AI replacing medical professionals, we are already in an era where many businesses are using generative AI (GenAI) to create content for their blogs, websites, and emails. However, it's important to note that this application is strictly marketing copy. This has many people exploring the viability of using GenAI to create technical content.


In terms of content, GenAI currently has a wheelhouse where it is very effective: short B2C marketing copy. Many GenAI platforms, like ChatGPT, do an exceptionally good job when it comes to generating high-level snippets of information - the type of copy you would use for product descriptions, short web paragraphs, emails, etc.


However, in terms of technical content generation, GenAI currently (at the time this article was written) comes up short. At this point, using GenAI to create technical content can be risky. In general, GenAI cannot generate technical content accurately and elegantly enough to do a plug-and-play method like you would with short, marketing copy.


Perhaps more alarmingly, when GenAI does generate technical content, it may inadvertently leave easter eggs of incorrect or falsified information in the document. For example, the lawyer Steven Schwartz used GenAI to write a legal brief which he then submitted after proofreading, but not vetting, the content. In the brief, AI referenced judicial opinions and legal citations that it made up 1. The results were mortifying for Schwartz and had a catastrophic impact on his career.


Schwartz is now known as the “ChatGPT Lawyer” - a title he will never outlive and that will impact his career for the remainder of his life. While Schwartz was the unfortunate first to become a cautionary tale, the fact is that this could happen to anyone using AI to generate content of any kind. Once you remove the human element from the writing, you introduce the potential for technology to fill in the gaps when it doesn’t have the information that it needs.


When humans don’t have the information they need, they go out and find it; when ChatGPT doesn’t have the information it needs…it makes it up. This approach in any content is problematic, but it is exceptionally problematic for technical content.


While it may seem like reviewing the content should be sufficient to remedy this situation, that’s not necessarily the case. When reviewing AI-generated content, you have to fact-check everything, ensuring that every claim and all the data referenced are, in fact, truthful, relevant, and authentic.


Rather than expediting the content generation process, you are introducing the risk for falsified information as well as increasing the amount of time you will spend validating the information. This means extra research, vetting, cross-referencing, and potentially even looking for sources that don’t actually exist (if Schwartz’s brief is any example to judge by).

Moreover, when GenAI creates longer-form content (technical and non-technical alike), AI will often lose the plot and begin repeating information within the first three paragraphs. This means that your team will still have to review and rewrite a significant portion of the final output from GenAI.


GenAI certainly has a place in content generation; however, it is not yet a magic bullet for content creation - particularly in the technical realm. Much like heavy machinery, it can be a very effective tool, but it needs to be operated by a professional. Businesses should be careful in how they utilize AI to generate content, and as of right now, we would recommend avoiding GenAI altogether when it comes to generating technical content. No one wants to be the “ChatGPT Lawyer.”


That said, every day is a new future with AI, and advancements are happening alarmingly quickly. Portions of this article could be obsolete within two months, five months, or a year. We're all just along for the ride.


If the ChatGPT Lawyer fiasco has convinced you that using GenAI to write your technical content may not be the best approach, reach out to us. We can help you generate the content you need to lead your industry in thought leadership and drive your audience to interact with your brand.


BONUS CONTENT: If you're interested in a broader look at the future of AI, Trevor Noah interviewed Sam Altman (the CEO of OpenAI who was recently fired from his position by the board of directors and then abruptly reinstated after a dramatic outcry from the OpenAI staff) on his podcast What Now.


Sources:

41 views0 comments

Recent Posts

See All

Comments


bottom of page