artificial intelligence fine-tuned from GPT-3.5, a language model trained to produce text. ChatGPT was optimized for dialogue by using Reinforcement Learning with Human Feedback (R.L.H.F.) – a method that uses human demonstrations to guide the model toward desired behavior.

The LISI team would tell you that one of my favorite sayings is “Let the robots do it,” and while I do like to automate boring tasks such as pulling data from multiple sources, I wouldn’t go so far as to say “let the robots do my job.” 

Content is a very important aspect of any digital marketing strategy, and we know it takes time, thought, and effort to create, as well as the expertise to share it. ChatGPT may help you with content creation, but before you start delegating blog posts to it, you may want to consider a few things. 

What are ChatGPT’s limitations?

To learn more about ChatGPT, let’s look at the tool’s documentation that details how the tool was trained and how it functions. A couple of items stick out as it relates to creating content:

Can I trust that the A.I. is telling me the truth?

ChatGPT is not connected to the internet, and it can occasionally produce incorrect answers. It has limited knowledge of the world and events after 2021 and may also occasionally produce harmful instructions or biased content.

We’d recommend checking whether responses from the model are accurate or not. If you find an answer is incorrect, please provide that feedback by using the “Thumbs Down” button. (1)

The content you produce as a firm should be truthful and represent your firm to the best of its abilities. So you need to be producing not only accurate information but timely information. The documentation tells us not to trust ChatGPT to always produce accurate answers, and to be aware of the timeframe. Suppose you are writing content around the TN Information Act, introduced in January 2023. ChatGPT would have no knowledge of this. Considering that you often want content to be fresh and on relevant timely topics, ChatGPT may not be the ghostwriter you want it to be.

Takeaway: Have a skilled (human) writer. 

Is the information I input into ChatGPT protected?

When you use ChatGPT, you will need to enter information into it to explain what you want it to write about. For example, you’ll need to tell the tool what the content is about, what tone of voice you want to use, and keywords.

Attorneys and legal marketers take note: ChatGPT will keep your data.

From the ChatGPT documentation: 

No, we [OpenAI] are not able to delete specific prompts from your history. Please don’t share any sensitive information in your conversations.

Yes. Your conversations may be reviewed by our A.I. trainers to improve our systems.

ChatGPT will occasionally make up facts or “hallucinate” outputs. 

If your law firm uses the tool to create content, these three statements should give you pause. ChatGPT will keep your data, so if you are creating content that involves real facts or cases, you don’t want OpenAI to have sensitive legal information. In addition, you would not want an A.I. to know your strategy or viewpoint as it relates to a specific case. Finally, much like the concern we highlighted above, ChatGPT could just be flat-out wrong. Again, use ChatGPT as a starting point, making sure a human reviews anything it creates. 

What Does Google Say About ChatGPT?

Google is currently the authority on content guidelines as it relates to S.E.O. They say what will rank well and what won’t. The following language, applicable to ChatG.P.T., had been in place long before the arrival of ChatGPT:

Spammy automatically-generated content (2)

Spammy automatically generated (or “auto-generated”) content is content that’s been generated programmatically without producing anything original or adding sufficient value; instead, it’s been generated for the primary purpose of manipulating search rankings and not helping users. Examples of spammy auto-generated content include:

– Text that makes no sense to the reader but contains search keywords

– Text translated by an automated tool without human review or curation before publishing

– Text generated through automated processes without regard for quality or user experience

– Text generated using automated synonymizing, paraphrasing, or obfuscation techniques

– Text generated from scraping feeds or search results

– Stitching or combining content from different web pages without adding sufficient value

If we unpack that rule a bit we can infer a few things:

 …..content that’s been generated programmatically without producing anything original or adding sufficient value

Google is always looking for high-quality content. Simply using an output from ChatGPT will likely yield less quality results than doing the research yourself and providing value and expertise on whatever topic you are writing about. In fact expertise, authoritativeness, and trust is another factor that can impact Google rankings. There is some debate as to whether E-A-T (Expertise, Authoritativeness, and Trustworthiness) is actually a ranking factor but we will get into that another time.

Text translated by an automated tool without human review or curation before publishing

Remember what I said about using Chat.G.P.T. as a guideline. I think it is perfectly acceptable to use A.I. to generate an outline for a content piece to speed up the content creation process, but it is crucial to have a human writer or editor to take that outline and write something. 

Text generated through automated processes without regard for quality or user experience

Simply copying and pasting from an A.I. content tool will not provide the best user experience for your readers. For instance, the A.I. may not understand context or acronyms that are common in the legal industry, thus the output may be confusing to readers if taken as is. More importantly, Chat.G.P.T. on its own cannot provide insight and opinions that give your content unique value. Google is looking for content that is useful to people. Your firm’s attorneys are subject matter experts in their areas of law and can add immense value to your content. (3)

John Mueller of Google didn’t shut the door on A.I. generated content forever, recently stating that at some point Google will focus “more on the quality rather than how it was generated.”

Takeaway: Use ChatGPT as a tool to speed up the content creation process by creating content outlines. Always have a human review the content any A.I. tool generates to make sure that it makes sense for your audience.

ChatGPT is a really cool tool, and I think it will help speed up the content creation process. I don’t think it will replace your lawyers, and certainly can’t replace their knowledge. 

(1) “ChatGPT FAQ.” Help.openai.com, help.openai.com/en/articles/6783457-chatgpt-faq.
(2) “Spam Policies for Google Web Search | Google Search Central | Documentation.” Google Developers, developers.google.com/search/docs/essentials/spam-policies?visit_id=638101820482937049-184081739&rd=1#spammy-automatically-generated-content. Accessed 25 Jan. 2023.
(3) Schwartz, Barry. “Google: Machine Written Content May Be Okay for Ranking Soon One Day.” Seroundtable.com, 8 Nov. 2021, www.seroundtable.com/google-machine-written-content-okay-32378.html. Accessed 25 Jan. 2023.