It’s been a while since I wrote anything here, so I thought I’d turn to artificial intelligence for some help. I decided to do this after a conversation with David King, one of my business partners at Tangilla. We were debating the meaning of “artificial intelligence.” For the past several years, I have used the following definition of artificial intelligence in my presentations. “Artificial Intelligence, broadly, is human intelligence displayed by machines.” The earliest uses were in games like checkers. And, by that definition, almost everything your phone does today is a form of artificial intelligence.
Further, in my view, “machine learning” is a method for achieving artificial intelligence and “deep learning” is a technique for implementing machine learning. With deep learning, like that employed by DeepMind or Deep Blue, the computer begins to mimic the human brain.
David’s contention was that we should not be using the artificial intelligence label for things that are “just” machine learning with narrow implementations. He feels strongly that the artificial intelligence label should be reserved for only those technologies that pass the Turing test. I’m dramatically oversimplifying our conversation, but I think I’m paraphrasing his position properly.
On the plane ride back from Argentina, I decided to demonstrate how far some of this technology has advanced. I’m going to be using only natural language processing software that is commercially accessible right now. Whether you decide to call this AI or machine learning is up to you.
The vast majority of the writing that comes after the subhead below was NOT written by me. I influenced it. As you will see from the attached video, the majority of the writing was done by two different pieces of NLP software. I decided to record the full writing session and leave it in real-time. It is located at the end of this blog post. The entire process for the following writing (roughly 900 words) took a little under 17 minutes. I have done no further editing after the video except to add relevant hyperlinks to the copy written in the process. Enjoy. 🙂
AI-Generated Content below:
Do you ever feel like you’re spread too thin – like you can’t possibly get everything done that you need to? That’s how I’ve been feeling lately, so I decided to do something about it and get myself a new writing assistant. Artificial intelligence has come a long way in recent years, and there are now AI programs that can help writers with everything from drafting essays to proofreading manuscripts. I’ve been testing out a couple of such programs for a while now, and I have to say that I’m really impressed! If you’re curious to learn more about this technology, keep reading – I’ll be discussing my experience with AI writing assistance in further detail below.
For this post, I’m combining two such programs. The paragraph above was predominantly written by Jasper.ai. I’m going to prompt the second AI writing assistant, Inferkit.com, with my own words. In this example, I’m going to use it as a sort of “muse.” I’ll stop writing myself at points and have Inferkit generate the next few lines of copy.
My hope is that this writing assistant will create paragraphs that are relevant to what I’ve been writing. Although I didn’t write the last sentence myself, I have tried this a few times in the past and I’ve found that AI writing assistants can sometimes help me fill in the gaps. I’ll begin writing, and in a few sentences, Inferkit will pick up the thread and begin filling in paragraphs. At first, this section may seem completely random and unrelated to what I wrote earlier, but with time, the AI writing assistant will begin to fill in the gaps.
I’ll be posting a video of the entire process and you will see that I’ve set Inferkit to the “blog” category and for it to start at the beginning in analyzing my writing to complete the sentences. If I keep writing, the process will continue over and over until I decide to draw this exercise to a conclusion. If you’re at this point in the video, I’m sure you’ll see where I allowed the AI writing assistant to take over and where I’ve written myself.
How does the AI writing assistant do this? The way AI works is that it analyzes the sentences you’ve written and detects patterns and patterns of sentences to write by itself, effectively completing the sentences that you’ve written. The AI uses deep learning, which means that the programs can analyze thousands of sentences and still learn new things.
As with any new technology, the AI is susceptible to glitches and glitches will inevitably show up in some of the suggestions the AI makes. This is part of the reason why I’m doing this little writing experiment. In future attempts, I’m hopeful the AI assistant can provide paragraphs that are better suited to the blog. What does this mean? At first, this may appear a bit odd to you, but when you see the finished paragraphs, you’ll understand that they are now tailored to the posts, based on my edits.
According to the creators of the AI writing assistant, it’s designed to be effective for writing blogs, essays, letters, and other documents for academic or business settings. In short, if you write for academic purposes, you’ll find that this AI can provide you with paragraphs that would have taken you quite a long time to write on your own. Or perhaps, the AI can simply act as a prompt to get you “unstuck” when thinking about what words to use next.
For example, I’m not 100% sure where I’d like this post to go from here or how I want it to end. When I hit “generate text” the AI assistant will generate the paragraphs and it won’t be entirely clear whether I want the AI to complete the sentence or not. But with practice, the AI writing assistant will probably provide me with the next few sentences that will complete the idea and you may start to see where I’ve had this post in mind for the past few weeks and where I’m simply writing it from the prompt of the AI writing assistant.
As you will see in the accompanying video, I used the AI text generation process in the paragraph above, but I rejected the first prompt in favor of the second prompt. This is where I think the main benefit of an AI writing assist lies today. When I get stuck, the AI assistant may suggest the next sentence to write to complete the thought. Or perhaps the AI can simply act as a prompt to get me “unstuck” when thinking about what words to use next.
And, that is exactly what the AI assistant has done in this case. I’ll come back to this process again in the near future and take the process in an entirely new direction. In the meantime, I highly suggest that you give a couple of these AI writing assistants a try. I think you’ll be surprised at the things this AI is capable of producing.
See my video below for more details and, if you have questions, don’t be afraid the leave them in the comments. 🙂
[…] excited to continue to introduce you to the new AI writing assistant I’ve added to my content creation process. These new AI tools help me write more content, […]
[…] that great writing? No. Not really. But it is certainly a fairly impressive display of what I would define as artificial intelligence. Might it be an interesting start to something or as a prompt for my own writing? Why not? With a […]
[…] of writing at scale with artificial intelligence. I already see Jasper.ai as a very capable AI writing assistant on a personal level. In that context, the webinar was dedicated to illustrating the power of what […]
[…] go. I lead it. That said, the rest of the writing that follows this sentence was also entirely written by AI based on a single prompt, “Is artificial intelligence just a very clever prank?” For […]
[…] been attempting to do some light illustrations of artificial intelligence writing assistants here on this blog for the last couple of weeks. My goal was to show what was possible with limited […]