I have several peers who work in the computer industry, including software programmers and IT professionals at various levels, ranging from small corporations to the public sector. When we get together, many of our conversations eventually turn to AI and its impact on our jobs and professions. The same thought comes up again and again:
Can AI be trusted for work that requires specialized knowledge? And if so, how?
A bold claim from a programmer
One of my peers recently confided that he is hearing about software companies insisting their programmers use AI for the bulk of their work. As an outsider looking in, my first thought was that this was a great idea. We frequently use ChatGPT in our process. With our specific direction, it helps lay the groundwork for the page, which frees us to add context and our client’s voice. We are able to create high-quality content in less time, giving our clients an edge over competitors who may be targeting the same market.
However, there’s a second side to his tale. To further reduce costs, these same software companies are also hiring inexperienced programmers. They mistakenly believe that AI can compensate for the skill gap they lose by replacing experienced programmers with inexperienced ones.
That’s a mistake. During my time using AI as a productivity tool, I’ve seen it over and over again; AI requires human expertise to be used effectively.
The common law of business balance in the digital age
Every business has a duty to find ways to remain lean and competitive. If there’s a more economical alternative, it should be explored. But there’s a caveat. Your cost savings should never diminish the quality of your product or your services.
When I hear stories or comments about businesses and people focusing solely on cost, I can’t help but think of the quote attributed to John Ruskin:
“It’s unwise to pay too much, but it’s worse to pay too little. When you pay too much, you lose a little money — that is all.
When you pay too little, you sometimes lose everything,
because the thing you bought was incapable of doing the thing it was bought to do. The common law of business balance prohibits paying a little and getting a lot — it can’t be done. If you deal with the lowest bidder, it is well to add something for the risk you run, and if you do that you will have enough to pay for something better.”
There is nothing wrong with wanting to save money, but if you value only cost over quality, you will lose money in the long run.
AI without experienced oversight is a disaster waiting to happen
ChatGPT and other AI-based programs and services are useful tools. However, you cannot trust them to replace the job of a trained professional.
You can have ChatGPT draft your content, and Grammarly check your spelling, but if you skip working with an experienced editor or writer who understands digital marketing and the legal industry, is adept at fact-checking, and knows how to engage readers, it may harm your website’s performance in search results. AI-generated content that is poorly written, inaccurate, or unable to match user intent will lower your website’s reach.
I have seen ChatGPT make every mistake imaginable, from misspelling the client’s name to giving false information on statutes. I have fed custom GPTs strict adherence guidelines and watched it deviate over and over again.
And then there’s the flat, repetitive message it produces. AI-generated content lacks any humanity or personality. It will follow the same structure and format every time, never conveying your voice or connecting with your clients. At best, you’ll be left with generic slop.
Hurdles like these can be avoided when utilizing experienced editors and writers who understand your market and can direct (i.e., prompt) the AI to produce content that’s uniquely you, and then edit and revise accordingly so your content is helpful, insightful, and based on your experiences and perspective.
Don’t trust something that lies to you
Anyone who works with AI like ChatGPT will tell you that, eventually, it will lie. When AI lies, we call it hallucinating. The LLM has massive amounts of data, not all of which is up to date or even accurate. When it generates content, it’s pulling data from multiple sources and giving the most probable answer rather than verifying an accurate one.
In essence, it’s programmed to hallucinate. It wouldn’t work if it couldn’t hallucinate.
“The technology behind generative AI tools isn’t designed to differentiate between what’s true and what’s not true. Even if generative AI models were trained solely on accurate data, their generative nature would mean they could still produce new, potentially inaccurate content by combining patterns in unexpected ways. Weise & Metz, 2023”
We only recognize those hallucinations when the content generated is false or erroneous, but your target audience may not spot these errors. Inaccurate content can cause serious harm to your company’s brand. In turn, your content will fail to connect with new clients, as search engines like Google prioritize relevancy and accuracy based on how users interact with flawed and generic content.
It’s bad enough to misspell a word, but imagine how bad it would be for a firm to post a page that got the law wrong.
AI works best when the user has a solid foundation
When generating content to market a law firm or promote a cooking channel, we need to know three things:
Our clients’ unique qualities
No matter how effective or efficient AI is, if we don’t know who our client is and what they stand for, the content will fail to connect with readers because it’ll be the same as everybody else’s content. Beyond the basics of the name and phone number, the content should convey our client’s unique selling points and what sets them apart from their competitors. We build client profiles based on:
- Awards, accolades, and accomplishments they have received
- Organizations and forums they belong to
- Charities and community programs they donate to
- Principles, ideals, and values they believe in
We also seek to answer such questions as:
- How long have they been in business?
- Where are they located and where do they practice?
- Why do they practice law?
- What inspires them to do what they do?
- Do they have a slogan, message, or tagline?
- What is their mission statement?
- Do they have any memorable cases?
- Do they have any preferences and dislikes (in their content)?
- How successful have they been?
- What do people say about them?
- How do they stand out from their competitors?
- How are they involved in their community? In their profession?
The more we know about their insights, experiences, anecdotes, and perspectives, the more powerful, rich, and genuine their content will be.
We need to understand their products and services
When writing content for a client, we must understand the basics of their industry, including their services and products. The more we know, the better equipped we will be to present professional, well-crafted content.
For a law firm, we need to know what types of law our clients practice. Do they practice personal injury, criminal defense, immigration, or something wholly unique? But more importantly, we need to know how that area of law works. We need to know what kind of clients they want to attract. If the content is too generic or targets the wrong audience, it produces unwanted leads. Additionally, we know how to read a statute and where to find them. We are also empathetic enough to understand what the reader is going through and how our client, the lawyer, can help. Finally, we recognize and adhere to state bar marketing rules, something AI could easily get wrong.
Now, Google and ChatGPT are great at providing helpful information to include in your content. However, you need the expertise to fact-check, spot errors, and develop a marketing strategy that elevates the company’s brand. This is where a company can truly save money: when it relies on experienced specialists.
Training in the fundamentals of SEO
Since I started working at We Do Web, the company has consistently encouraged our growth in SEO skills and knowledge. Each week, we read and review articles offering insights, opinions, and updates on the SEO industry. And every time I read an article, I find that the fundamentals for generating content remain the same. They may change the tune, but it’s still the same dance.
When we generate content, we create content that shows:
- Expertise: Our client’s knowledge and skills
- Experience: Our client’s time practicing, their successes, and lessons from lived experiences
- Authority: Our client’s reputation in their field and online
- Trustworthiness: How the community views our client
Our goal is to create content that showcases the best qualities of our client. But we also need to execute the basics, like:
- How to build a strategy that earns visibility for high-value keywords and user prompts
- Linking to reliable sources with proper anchor text to demonstrate authority
- Addressing the reader’s intent and needs so that we deliver the helpful content that solves their problem
With the right inputs and instructions, ChatGPT can get you started, but it takes an experienced writer to make your content shine.
Get what you want or get what you pay for?
Maybe it was a disservice to define ChatGPT as an “AI.” Perhaps it gave the public the wrong impression that it was an autonomous, thinking entity. We still have a long way to go before we can have that sci-fi experience of a computer asking us, “Please state the nature of the medical emergency.”
Relying on AI to generate a product unsupervised is a disastrous mistake. We can’t even rely on ChatGPT to accurately count the number of words on a page. If you choose to use AI without careful oversight, you should set aside funds to correct any errors and inaccuracies it may introduce.
The more economical choice is to rely on experts to make a quality product.
