Somewhere between your holiday festivities, travel mishaps, or planning your New Year resolutions, you may have caught wind of ChatGPT. This new and innovative AI platform has set the internet abuzz just days after being released for public testing on November 30, 2022.
Among its various capabilities, ChatGPT can generate content based on prompts, similar to extremely advanced predictive texting. It can also answer questions, compose music, create graphics and animation, generate copy for Google Ads, synthesize complex information, and even write novels. The quality of the content it produces can be surprisingly good, even for AI. But what do ChatGPT and its capabilities mean for legal content writing and law firm marketing?
First, let’s dig into what ChatGPT is. ChatGPT is a chatbot form of artificial intelligence (AI). It’s a variant of GPT-3 (Generative Pretrained Transformer 3), a state-of-the-art natural language processing model. It answers complex questions conversationally as it is trained to understand what humans mean when asking a question.
With directions, it can accomplish many different tasks, such as creating outlines, writing code, or even drafting social media posts. Programmed with significant amounts of data, ChatGPT can predict the next logical word in a sentence.
The more data it receives, the more it can accomplish—much like a human being. With these capabilities, it’s not surprising that businesses are exploring ways in which ChatGPT can maximize the effectiveness of their marketing budget.
However, as with all large language models, it isn’t trained to do particular tasks. With its wide variety of knowledge, it can perform tasks it hasn’t been specifically programmed to do. For example, a human can be given knowledge and taught to bake. They don’t need specific knowledge on how to bake a chocolate cake, only general baking knowledge.
Similarly, ChatGPT can be trained with information about general topics and apply it to specific tasks. In addition, it can be given directions and various constraints—such as a word count and specific talking points to create content. Furthermore, there are certain safeguards in place that will ensure that the model refuses inappropriate requests.
ChatGPT is the brainchild of OpenAI, a non-profit artificial intelligence research company founded in 2015 by CEO Sam Altman, Elon Musk, and several other Silicon Valley investors. OpenAI changed to a “capped-profit” company in 2015.
In 2018, Musk stepped down from the board because of a conflict of interest between OpenAI and Tesla’s autonomous driving research. However, Musk is still an OpenAI investor and expressed enthusiasm for ChatGPT’s launch. According to Altman, the software reached one million users less than a week after its launch.
Since ChatGPT can be used to create virtually any type of content, it can hypothetically be used to create legal content. With its vast knowledge and predictive sentence building, all it needs is instructions from you about what topics or practice areas to write about, what you would like it to include, and how long it should be. The results will likely be similar to the content that a human legal content writer creates.
However, just because you can do something and have the ability to do it doesn’t mean that it’s a good idea and that you should do it. This most certainly applies to using ChatGPT to create legal content. There are many concerns with using ChatGPT and other forms of AI to develop legal content that everyone in the legal marketing realm should be aware of.
At the time ChatGPT was released, Google considered AI-generated content spam. In the months following ChatGPT’s release, the search engine has clarified its position that it is not against AI content per se, but is against using AI to mass produce low-quality content in an effort to manipulate search rankings.
In March of 2024, Google released a core update in response to an increase in spam results in the search results. That update specifically targeted “scaled content abuse” – or the practice of generating content at scale to increase rankings – as a practice it would penalize. While Google claimed to be agnostic regarding how content was produced, many industry observers saw the update penalized that had AI -generated content. In fact, some sites that rely heavily on AI-generated content were completely deindexed from the results.
While Google’s official position is that it rewards helpful content regardless of how it is produced, it’s advisable to use AI with caution in your content creation process.
Despite being able to produce near-human content and answers, ChatGPT is not human. It can be wrong, sometimes dangerously wrong, leading those searching for answers astray. Consider the answer to this question:
While some of this information is correct, ChatGPT certainly got the answer to the question wrong. If your legal clients are searching for reliable information and someone they can trust, ChatGPT can’t be trusted to give them what they need.
ChatGPT isn’t trained for accuracy. According to Contently.net, OpenAI freely admitted in a blog post that “ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers. Fixing this issue is challenging, as (1) during RL training, there’s currently no source of truth; (2) training the model to be more cautious causes it to decline questions that it can answer correctly; and (3) supervised training misleads the model because the ideal answer depends on what the model knows, rather than what the human demonstrator knows,” the company wrote in a recent blog post.
OpenAI also states that ChatGPT can’t provide accurate information on topics and changes after its knowledge cutoff in 2021. If the information could only be found on the internet in 2022, it won’t be included in ChatGPT results. So, any new information or changes to the law won’t be factual when using ChatGPT.
ChatGPT is a dialogue system. As such, it’s usually biased by what you previously asked it. This context bias and its statistical nature can result in different solutions to the same problem, some of which may be incorrect. Additionally, it generates answers by making a series of guesses, in part causing it to argue wrong answers as if they were entirely true. Furthermore, it has been shown to exhibit biased behavior.
While getting the wrong information about elephant reproduction may not be highly problematic, legal or factual errors in your online content can result in bar complaints or worse. All state bar associations have rules of professional conduct that lawyers and law firms must comply with. Most address that lawyers can’t engage in deceptive or false communication about the law or their services.
Whether you or AI created the content you post, you are still responsible for it. For example, suppose you post inaccurate content about your firm, services, or the law. In that case, you could face penalties under your state bar, up to and including being disbarred. Leaving your legal content creation to ChatGPT or a similar AI platform could be disastrous for your practice or your career.
In addition, providing inaccurate information on Google will cause your content to rank lower. If your content is incorrect, it decreases user and Google trust and your E-E-A-T (see below). You can’t expect to rank on the first page of search results with inaccurate information on important topics like legal matters. Using ChatGPT can defeat the purpose of even posting content online in the first place.
ChatGPT discourages its use for creating legal content and advice due to problematic model outputs. Consider what ChatGPT itself says about providing legal advice:
OpenAI wants to make it harder for humans to pass AI off as their own work. Cryptographic watermarking technology has already been created to help detect content created through AI, including ChatGPT content. While ChatGPT content isn’t currently watermarked or secretly identified in any way that we know of, OpenAI states that watermarking is “hopefully” coming with the next version.
With cryptographic watermarking, anyone with a key can test a document or content to detect if there is a digital watermark present that establishes the content was indeed generated with AI. This watermarking would generate a long text or an undetectable secret sign that the content was produced using GPT.
Not only would this decrease spam content online, but it would also help prevent academic plagiarism and mass-generated propaganda. In addition, watermarking AI-generated content will allow Google to recognize and penalize sites that use it to try and manipulate the search results.
There’s also the potential that ChatGPT could use copyrighted materials in the answers it provides, which leads to its own Pandora’s box in the legal realm. Who owns the copyright? Can website owners get into legal hot water for posting copyrighted content they gleaned from AI? Staying away from ChatGPT and other types of AI or only using them as tools to aid in writing and SEO are the only ways to protect against what issues might come.
If you play around with ChatGPT, you will see that the content it creates is often extremely similar, even when you use different prompts. As a result, posting content created by ChatGPT can put you at risk of plagiarizing existing content or getting hit with a duplicate content penalty.
Finally, it’s critical to note that the legal profession is extremely slow to adopt new technology. In fact, even the ability to electronically file pleadings with the court or use electronic evidence is a relatively new development. In light of this technological reticence, it’s reasonable to assume that some state bars may prohibit lawyers and law firms from using AI-generated content in their marketing materials or work product.
On December 15, 2022, Google released its latest quality rater guidelines (QRG). Of note for your law firm website and marketing campaigns, these critical new guidelines include:
It should also be noted that Google itself is using more inclusive language, such as “website owners” instead of “webmasters,” and abandoning some gendered pronouns for “themselves.”
The QRG documents are essential for anyone who works in search marketing as they are a complete guidebook for the direction Google wants its algorithms to go. By reading between the lines of the QRG’s language, you can help determine what Google is looking for when it comes to content quality, user experience, and E-E-A-T of websites.
Achieving visibility in Google search is only accomplished by following the guidelines, including those for non-spammy, original content, and E-E-A-T.
You may already be familiar with Google’s E-A-T acronym, which stands for Expertise, Authority, and Trustworthiness. However, Google recently added a new E at the beginning of the acronym for Experience.
Law firms, content creators, and SEO specialists now need to ensure they include experience in their Y-M-Y-L content. They need to establish first-hand knowledge about the subject they are publishing content on. Going forward, content quality will also be evaluated by the extent of first-hand experience in the topic. In light of this guidance and the fact that ChatGPT can create content for people who lack experience in a certain area, it’s a safe bet that Google will be on the lookout for AI-generated content moving forward.
In addition to another E, Google also says that “trust” is at the center of this concept, and it’s the “most important member of the E-E-A-T family.” Importantly, content that is not trustworthy can have low E-E-A-T even if it demonstrates significant experience, expertise, and authority.
Google provides the following example it it’s search quality guidelines: “For example, a financial scam is untrustworthy, even if the content creator is a highly experienced and expert scammer who is considered the go-to on running scams!”
Experience makes you more trustworthy in the content you produce. Your site visitors and potential clients are more likely to trust you if you have experience in the area of law in which they are seeking help.
In March of 2024, Google updated its search rater guidelines in light of its contemporaneous core algorithm and spam policy updates The updates provide characteristics of untrustworthy pages as well as examples.
Specifically, the document indicates that a page should be considered untrustworthy if it has “multiple or significant factual inaccuracies on an informational page which would cause users to lose trust in the webpage as a reliable source of information.”
What does this mean for law firms? It’s more important than ever to ensure that all of the factual and legal information on your site is accurate and up to date. Regularly auditing your content to ensure that it is accurate will give you the best chances of ranking well in the search results.
Even though ChatGPT can be problematic when it comes to legal content creation, it and other large language models can be helpful tools. They can actually enhance the marketing efforts of most law firms when utilized correctly in the right space.
Your best use of AI tools like ChatGPT is for achieving SEO by making a worker more productive, not by replacing the work of a human. This usually means using AI to do the tedious work of research and analysis, which are some ways that you can target clients and leverage AI to improve content marketing:
By asking questions in ChatGPT, a law firm or legal content writer can find out what potential clients might be asking and the legal issues they want to read about. ChatGPT can help direct their efforts to produce valuable content on the law firm’s website. Connecting with a target audience must be done on their level.
If you are blogging about topics and answering questions your target client aren’t concerned with; you won’t connect with them. However, adding this tool to your content strategy makes you more likely to convert leads.
Creating compelling headlines can help with your SEO and help you connect with new clients. AI like ChatGPT can help you come up with those headlines after you provide it with specific instructions or parameters. For example, you could ask for “three great headlines for an article about joint child custody.”
Likewise, you can ask ChatGPT to create an outline for a 1000-word blog on “Best Defenses Against Drunk Driving in Illinois.” Of course, you will still want to review any results you receive, but AI can help you create and brainstorm the structure of your content.
ChatGPT can also help with law firm topics and keyword research in a law firm marketing plan, no matter the lafirm’s’s practice area. For instance, you ask ChatGPT for 50 search terms on pedestrian accident lawyers in Seattle. These kinds of prompts can be particularly helpful for local SEO. In fact it’s a valuable tool for conducting research as it can generate a large number of suggestions and specific keywords on any given topic. In addition, you can use it to generate both high search volume and long tail keywords, both of which can inform your content writing for landing pages and blogs.
You can rely on ChatGPT for law firm marketing or legal topics. However, you can leverage it to your advantage when humans interact with this AI technology. Whether you are a small law firm owner, an SEO specialist, or a legal content writer, there are ways that ChatGPT and other AIs can help you perform your job more efficiently.
No matter how good AI gets, many things in life still require a human touch. Think about times when you tried to use an automated chatbot for a customer service issue. It’s very likely that the chatbot couldn’t understand your specific problem. In those instances, it’s often better to call and speak with a customer service representative.
When seeing a doctor or a mental health professional, would you prefer to speak to a human or interact with a robot? Not only is it more personable and interactive to speak to a human, but they can also often give quicker, more realistic, and more trustworthy answers—which is what Google strives for anyway.
You can’t rely on ChatGPT to provide accurate answers or you might end up with the legal equivalent of an elephant laying eggs, which could prove disastrous to your career. Working with expert legal content marketers is how you can ensure that your content is accurate and compliant with the rules of professional conduct required for all law firms and attorneys.
For example, most state bars require attorneys to steer clear of using the term “specialty” or declaring themselves as an “expert” on legal matters. Since ChatGPT uses a large language model, it doesn’t know not to use this word when writing legal content for law firm websites. In addition, it doesn’t know where the line is between providing legal information and legal advice.
In a regulated profession like law or medicine, you need to be absolutely sure that your content is accurate. ChatGPT and similar forms of AI can be helpful to a law firm, but they require a human to leverage them correctly.
There’s no denying that AI has come a long way, especially with the onset of ChatGPT technology. This form of AI has significant real-world applications in content marketing that will likely alter the course of online marketing strategies. However, generating creative, original content is a uniquely human ability, and it is unlikely that AI will be able to completely replicate this skill in the near future.
If you’re selling espresso machines, by all means, copy and paste the ChatGPT-generated content into your ads. But on the other hand, if you’re a lawyer or a digital marketing agency that works with lawyers or law offices, think twice about relying on AI to create accurate, long-form legal content.
David is a 2009 graduate of the St. Louis University School of Law. He got his start in legal marketing at a now-defunct legal marketing agency as a content writer. It was during this experience David recognized the issues firms and agencies have sourcing quality legal content at scale and founded Lexicon Legal Content to solve these problems. Since 2012, Lexicon has been providing firms and marketing agencies with accurate, optimized, and ethics-compliant content across a variety of practice areas and jurisdictions.
