Part 2 – Open AI and how content creators, like me, really feel about it.
Table of Contents
Spotting Generative Content in the Wild
Should AI-generated content be labeled
This article specifically talks about finding your voice when using AI for content creation. All images are Generative.
I swore I would follow up on my last blog post about Chat GPT and here’s the deal. I’m so obsessed with AI right now that it’s all I can think about, so this follow-up post was easy. Everyone is immersed in AI right now. That is a fact. If you are in business, school, etc. you are 100% using AI for something. Grammarly, Bing!, Zoom, Chat GPT, Adobe, Google whatever it is AI is sliding right into the mix. It’s a suggestion here, a slight rewrite there, and of course, you had to be living under a rock not to see the U2 concert in Vegas at the AI venue of all venues, Sphere. You know what, in most cases, the robot is a pretty nice little companion in my work day. It helps immensely to speed up the monotony of the daily mundane necessary to complete work these days.
I for one am enjoying it, but I’ve also seen enough Sci-Fi in my life to identify with the existential threat that AI experts are warning of. That’s another opinion for another day. For now, I’m all for letting the legislators and policymakers sort that out. What I do cringe about right now is something I hope to educate my own customers on. The misuse of Chat GPT in your marketing. At the end of the day, AI can be very helpful, but we’re encouraging you to think of it as more of a brainstorming tool than anything else.
Chat GPT in the hands of the customer is easy to spot
First of all, ‘Hands of the customer’, to us, means that you aren’t learning to become a prompt engineer anytime soon. You’re likely using Chat GPT Free and you’re likely using it in the same manner you would a search engine like Google.
Below are a few of the dead giveaways in the articles that I have been receiving:
- The content always has a title that reads Introduction and Conclusion.
- There is almost always a colon in the headline of the article
- There will be a list and it will probably have emojis
- The introduction and conclusion are almost identical
- The content is redundant, especially in the middle where readers tend to lose interest in articles
- The content is mildly incorrect and can contain everything from grammatical errors to factual inaccuracies and relevance issues.
Secondly, I still struggle with whether or not being able to identify that something is AI generated is bad. I’ve seen more customers take an interest in their own content for once in my life. Is that a bad thing? A lot of people struggle with the mechanism of writing, but they might have the imagination and ideas. AI can help immensely for those less educated, equipped, or inclined to bring their ideas to life.
Should creative, news, and anything that AI touches be labeled?
I don’t see the harm in that – yet. I don’t think that just because content has come from a machine it’s automatically fake or disinformation and doesn’t add value. AI is going to reshape a lot of industries, especially when it comes to education. The way we learn has been fundamentally changed and in ten years it will be a moonscape compared to what we’re used to. Forget new grading systems – that’s not worth complaining about.
For example, my daughter was discussing a friend who is in the Engineering department at her University. He had a paper to write. He used Chat GPT to formulate the topic for his paper and generate the base of content for it. After he got it to where he wanted it, he moved it over to the AI Assistant in Grammarly. Grammarly made some further suggestions and edits. Essentially, it allowed him to polish Chat GPT’s grammar and imbue more of the tone he wanted. The whole process took less than an hour. His weekend was free. Did he learn anything? I can’t say, and maybe that’s where the issues will stem because there are some jobs that robots can’t do – yet.
Teaching AI my voice – not as easy as I hoped
A big thing that younger students and those with less experience in the computer science department are missing is the ability to create meaningful prompts. Rather, they do as I stated above, and use Chat GPT the same way as a search engine to produce a pretty generic response. Write me an article about mortgage rates, why are high-interest rates bad, etc, etc. I was really interested in trying to teach the AI how to learn my voice. I found an article on Forbes that laid out a five-step process and the results were not quite what I wanted, to be honest. For the purpose of this article though, they’re nothing short of amazing to show the difference between human writing and AI writing even with some elevated prompts, training, and a background in “the industry”.
Take a look below.
AI for content creation – The first prompt is the deepest
I started with my last article as a jumping-off point. My goal was to use AI for content creation to generate a follow-up article that sounded like I genuinely wrote it. I began by using the Ghostwriter prompt and essentially told Chat GPT to analyze my writing style and tone. Below is the prompt I used and the subsequent AI response.
You are an expert ghostwriter, world-class at capturing your client’s authorial voice. You are also an expert in natural language processing. Below is a blog post written by your client. Please use natural language processing to create a paragraph that describes key characteristics of your client’s voice, so that an LLM could write in a similar voice using only the paragraph as input: [pasted content from https://atomic55.net/open-ai-and-how-content-creators-like-me-really-feel-about-it/ ]
How Chat GPT interpreted my voice
Your client’s authorial voice in this blog post can be described as informal, conversational, and introspective. They start by addressing the prevalence of AI and OpenAI, expressing a certain weariness about the constant discussions surrounding the topic. They use colloquial language and rhetorical questions to engage the reader and convey a sense of relatability.
Your client’s voice reflects skepticism initially, with a touch of resistance to the idea of AI and Chat GPT replacing human creativity. They narrate personal experiences and conversations, including heated debates with family members, which adds a personal and relatable touch to the writing.
As the blog post progresses, your client’s voice becomes more contemplative and self-reflective. They discuss their observations and concerns about AI, both in terms of its impact on education and content creation. There’s a sense of curiosity and openness to exploring AI’s potential benefits, even though it comes with a hint of guilt and uncertainty.
Overall, your client’s voice is characterized by a balance of skepticism, curiosity, and a genuine desire to understand and navigate the evolving landscape of AI and Chat GPT. It combines personal anecdotes, introspection, and informal language to connect with the reader and convey their evolving perspective on the subject. An LLM can emulate this voice by using a similar tone, rhetorical questions, personal anecdotes, and a conversational style to address the topic of AI and Chat GPT.
I found the analysis of my writing voice to be bang-on accurate. So much so, that I had high hopes for the next prompt using AI for content creation.
The second prompt – designed for a better you.
You are still an expert ghostwriter. You already have a natural language processing analysis describing the client’s authorial voice. The client has indicated that they also like Tina Fey’s voice and character, so please mix in 15% of Tina Fey’s writing style. Can you please write a blog, to update readers of the last post about how the client has used Ai and Chat GPT since the last posting? The blog post should contain the following facts: the client does not use AI for reports, emails, or any personal communications with other clients. The client is using it more frequently to formulate articles and routine work, however, they find it very redundant and that it needs a lot of editing and polish to make the writing appear more natural. The client has identified key nuances and is able to easily spot the use of AI in everything from news articles, headlines, social media posts, and other clients sending in blog articles for posting. The client has also started using generative AI for images and is in love with this process. The client sees the benefit of originality and uniqueness when looking at generative AI for artistic purposes. The client states that they don’t feel guilty about this because it requires imagination to generate the ideas – they’re just able to materialize much faster. The client’s daughter has been using generative AI starting with her final project last year. This year in university she is considerably further immersed in Open AI, Machine learning etc.
This is the article Chat GPT returned back to me.
Embracing the AI Wave: An Update on My Journey with Chat GPT
Hey there, dear readers! It’s been a while since I last shared my thoughts on the ever-evolving world of AI and Chat GPT. Today, I’m excited to give you a peek into what’s been happening since my last post. Plus, I’m adding a sprinkle of Tina Fey’s witty charm to keep things entertaining.
First things first, let’s talk about how I’ve been using AI lately. Drumroll, please… I still refuse to let AI infiltrate my reports, emails, or any personal communications with my clients. Call me old-school, but some things are sacred, and I believe the human touch is irreplaceable in these areas.
Now, onto the juicy stuff. I’ve been experimenting more with AI, especially when it comes to crafting articles and handling routine tasks. But, here’s the kicker: I find it as redundant as rehearsing lines for a comedy sketch. Sure, it helps with the heavy lifting, but boy, does it need some serious editing and polishing to make it sound remotely human.
Speaking of edits, my AI radar has been on high alert lately. I’ve become Sherlock Holmes when it comes to spotting AI-generated content. Whether it’s news articles, catchy headlines, or even social media posts, I can almost smell the algorithms at work. It’s like having a superpower, but it also makes me appreciate authentic, human-generated content even more.
Now, let’s dive into a thrilling new chapter of my AI journey – generative AI for images. I have to admit; I’ve fallen head over heels for this creative powerhouse. It’s like having a magical art assistant, and Tina Fey would probably agree that having a sidekick like this is the ultimate dream.
You see, I’ve discovered the beauty of originality and uniqueness in the world of generative AI for artistic purposes. And here’s the twist – I don’t feel an ounce of guilt about it. Why, you ask? Because it takes imagination to fuel these AI engines. We’re not just replicating; we’re conjuring ideas at warp speed. So, I’m all for it. After all, Tina Fey herself would appreciate the efficiency of turning ideas into reality faster than you can say “Liz Lemon.”
Now, for a quick family update. My daughter, who’s been riding the AI wave with me, has dived even deeper into the world of Open AI, Machine Learning, and all that techy jazz. She started with generative AI for her final project last year, and now she’s swimming in it like a pro. It’s incredible to see how quickly the younger generation adapts to these innovations. I guess she takes after her old mom, huh?
In conclusion, my AI journey continues to be a rollercoaster of discovery and intrigue. While I remain a steadfast guardian of human communication, I can’t help but embrace the AI tools that make my professional life more efficient. Tina Fey’s sharp wit reminds me that humor and creativity can coexist with technology. So, here’s to the future, my dear readers, where AI and human imagination collide in a harmonious dance. Stay tuned for more updates because this journey is far from over!
AI for content creation – It didn’t pass
I found that even using a sophisticated, detailed prompt couldn’t keep Chat GPT from initially revealing itself. In fact, the very first thing on my list of key ways to determine if it’s machine-generated content or authentic content was the title. And there it was. A colon. Next, was the 15% Tina Fey direction. It couldn’t use her voice and style, instead, it used references to her throughout the entire article to tell readers, hey this is just like Tina Fey. This was just one area that made the content cringe-worthy.
However, there were still three more prompts to get through in order to train this follow-up piece to emulate my voice.
More on that in my next follow-up to the follow-up – Part 3 is coming.
