I watched a webinar put on by Jasper.ai yesterday that illustrated the very present potential 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 they call “recipes.” Jasper.ai’s recipes “are pre-built workflows that contain a series of Jasper commands to help you create content using a repeatable process.” You can watch a replay of the webinar here, How To Create Custom Jasper Recipes To Streamline Your Workflows With AI or below (which starts at the recipe for idea generation).
Automated Writing At Scale
The first recipe they demonstrated was an idea generator. It was fed three variables; number (number of ideas to generate), topic (what are the ideas for), and niche (what specific niche is being covered). Other recipes contained different sets of variables specific to the task being accomplished. In each case, however, the AI was more than capable of handling the writing tasks at a speed no human could possibly match.
I decided to try out the first recipe for myself. Without much thought, I provided the following variables; Number (10), Topic (blog post titles), and Niche (REALTOR Association Management Software). I was curious to find out how it would handle such a specific niche. And here was the output.
- Comparing the top real estate association management software
- The benefits of using a real estate association management software
- The top features of real estate association management software
- How to choose the right real estate association management software for your needs
- Getting the most out of your real estate association management software
- Tips and tricks for using real estate association management software
- The top mistakes to avoid when using real estate association management software
- How to make the most of your real estate association management software
- The top benefits of using real estate association management software
- What to look for in a real estate association management software
Considering that the entire process from setting variables to blog post title idea generation took less than 60 seconds, that’s not a bad beginning list. I could easily edit these for my specific purposes. I’ll likely go back and add some keywords to the process to see what comes out of the tuning.
And, while It was not demonstrated in the webinar, it’s clear to me that it would be possible to populate an even more sophisticated recipe with variable data at scale if they were to make that input available via API. After you watch the webinar yourself, I think the applications for something like that will be pretty clear. And there will be good and bad actors playing in this space, with or without their help.
Inferkit Batch Processing
Inferkit already provides a way to generate copy at scale via their “Batch Request” feature. Batch requests automate the process of generating many outputs, or “writing at scale.” You can use a single prompt or a CSV of many different prompts. Once the outputs are ready, you can download them in CSV format for use as you see fit.
I decided to give it a try with two different single prompts. You can see the unedited resulting CSV files here: Inferkit Batch Tests, but I am including the edited text output from each prompt below. I’ve edited out most of the incomplete end sentences and accepted some suggested changes from Grammarly for ease of reading. I think both results serve as a good illustration of the potential. I only altered the settings slightly for the second prompt, but the process was the same for both.
Prompt 1: The problem with AI writing assistants is easy to define. They are already too good. It’s scary. In a few years, they will
- be able to write better and faster than you, and that’s a problem. To bring the world to your product, humans are needed. That’s where humans get it wrong. If AI is too good, why should anyone care about human rights?
- be able to write better and better persuasive pieces. AI, and I say this as a former computer scientist, is way ahead of humans in writing quality articles on personal finance. Here is a preview of the future, courtesy of Wikipedia.
- write better than human writers. Forget the danger to democracy. The problem is that the creativity and craftsmanship of humans will be obsolete.
- be writing our reviews and our Wikipedia articles, judging us on our taste in books and TV and our humor and all our opinions. Our likes, dislikes, passions, and even our thoughts.
- beat us at every competition we enter. It doesn’t have to be this way. While many people consider the concept of artificial intelligence from a theoretical standpoint if they start to ask tough questions and delve into the possible applications…
- be writing novels that we can’t read. There’s no better example than this video from CBS News. What’s eerie about this is that, when the AI system was struggling to write a witty five-sentence summary of a book about writing, a human researcher was…
- write entire books, based on learning, rather than being given preset rules and instructions. What a day that will be. As we move into the world of AI-driven virtual assistants and the real world will change as well.
- have the ability to write our news stories, instead of us having to think up ones. They will be able to cook up stories, instead of picking up a pencil and a piece of paper.
- get more so. What we see today is not what we’ll get in a few more years. If you think there is a big gap between the artificial intelligence agents and you, think again. There’s a bigger gap between you and the agents.
- be able to write your best-selling novel or memoir. But I’m not talking about a feeble imitation. I’m talking about AI writing characters, emotions, and relationships – well, if not love – at least mutual respect, understanding, and compromise.
Prompt 2: When artificial intelligence is at it’s peak performance we will see
- the true potential of this virtual assistant becoming a necessity, not a luxury. We will be able to interact with our assistants just as we do with other humans.
- more world-beating talents. Think about all the space-based problem solving, like identifying planets or planets around other planets, to comprehend planetary physics, etc.. It will be limitless in its possibilities and that’s how you get a little glimpse into the future.
- deep-learning become an integral part of most aspects of our lives. The good news is, this is a new technology and there’s still a lot more to come in terms of its applications.
- more of an emphasis on simulations, and in a way, less of the real thing. We’ll probably also see more sophisticated predictive algorithms that serve up real-time predictions for what is going to happen, and when, in all of our complex social, economic, and legal situations.
- massive changes in the way organizations and governments do business. Not to mention how digital marketers will take advantage of this technology to engage customers in a whole new way. The application of digital marketing techniques and technologies to artificial intelligence is how you get the most bang for your buck.
- driverless cars, near real-time translation, the end of cash and online banking, world hunger eradication, and a better than human level of understanding of the human condition. Richard Branson, Elon Musk, Bill Gates, and nearly every other tech visionary in Silicon Valley, describes AI as the most important development of the…
- a hybrid system of physical humans. This should be kept in mind when we use the term human intelligence in its correct context. “Deep learning” is the hardest type of AI to build. It involves a lot of data and amounts to a staggering amount of computation that it takes to achieve good results.
- the online marketplace evolve to the next level, and shopping online will be as accessible as ordering from your favorite restaurant. AI can scan a thousand face pictures in seconds. It can learn what a smile means. It can comprehend what a raise means. It can understand if a face is sad or happy. It can add to a picture…
- the broad emergence of radical innovation, with the potential to disrupt entire sectors, according to Steve Omohundro, CEO, and co-founder of Omohundro AI. (Note: While Steve Omohundro is a real computer scientist specializing in AI, I can’t find any evidence that he actually said this. In addition, there is no company named “Omohundro AI” that I can locate. This is one of the dangers of this kind of transformer-based text generation. It just makes some things up on the fly.)
- the birth of “cyborgs”. That’s when humans become absorbed into machine-like beings, blending with technology, programming, and a whole host of different innovations, ultimately choosing the best features of each to form a hybrid entity with a brain and a body.
I see some obvious utility in tools like these. Even with Inferkit, if all I were looking for were ideas on different paths to follow with these prompts, the results above are more than adequate as either starting points or different takes on the same theme. And all of that took place in minutes with zero effort.
Anyword is another AI copywriting platform I’ve been playing with a bit. My first effort on the Anyword platform was a “Tweet Project” for Tangilla. I used their “rewrite engine” to present multiple new tweet ideas. I was pleased with the first tests, so I’ll keep playing.
In an email from Jenna Malatskey, Anyword’s Content Marketing Manager, she said this about the future of AI writing assistants: “The bottom line? AI writing tools aren’t replacing humans — they are helping humans with human problems. AI helps cure writer’s block. AI helps eliminate tedious tasks. AI helps copywriters improve marketing copy performance and allows them to do what they do better.”
And to the point I’m making here, this and tools like it will accomplish this and more with increasing scale much sooner than many may be willing to accept.