How Exactly Did Jasper Write That?

Print Friendly, PDF & Email

* Note: Jasper recently changed to its present name from Jarvis

As a techie, I enjoy reading industry journals and keeping up with the latest trends. However, the copywriter side of me was spooked recently when I came across something that was amazing but left me a little concerned for my future prospects.

No, it wasn’t a Bureau of Labor stat about the job prospects for freelance writers. Few of us are earning Stephen King levels of cash anyway.

The article I read was entirely created by Jasper, one of a growing batch of AI platforms that are generating reams of marketing content with almost zero human intervention. The implications were almost enough to leave me quivering in the corner, worrying that robots were coming for my job.

Well, not really. But it did leave me fascinated and wanting to learn more.

What Exactly is Jasper?

No, it isn’t the AI-powered assistant from Iron Man. In simple terms, Jasper is an artificial intelligence bot that uses things like Natural Language Generation (NLG), machine learning, and other forms of big data analysis to create highly tailored targeted marketing content. Jasper is already being used by companies like Logitech, Harper Collins, Zillow, and thousands of others to generate blog content and social media posts on a regular basis.

However, this company isn’t the only vendor to use such AI-powered technology for copywriting. It is available from dozens of other platforms and content marketing companies.

This tech isn’t even as new and untried as you’d expect, though it has leaped forward in quality and sophistication in the past few years. Unlike the Harry Potter that Botnik Studios released back in 2018, the Jasper-generated content I read was cogent, concise, and presented to hit every KPI marketers chase.

This particular tool is easy to use, and the results might be called amazing. Based on an open-source technology called GPT-3, coupled with minimal human input, copywriter bots can effortlessly create virtually unlimited content that’s more than 99 percent unique and error-free. Licensing for the underlying Codex model of GPT-3 was acquired by Microsoft in September of 2020, but the public can still use the API to receive output.

At the most basic level, users simply type in a general topic and tone and let the technology do the rest. There is a broad range of parameters and templates available to help create personalized, purpose-driven content that’s generated according to client input.

These apps are powered by Big Data analysis to uncover patterns, trends, and correlations found within the 74 zettabytes of data already in existence. They then use this exposure to learn about human speech patterns and create new copy based on the examples it studies. 

However, the exposure for each type of application must be rooted in the industry in which it’s used in order for it to be relevant for marketing purposes. An AI copy generator used for marketing must be exposed to marketing content. When using such technology for eCommerce, it should be exposed to existing examples of eCommerce-related content.

Otherwise, expect useless, disconnected copy.

I asked a friend, Zach Stein, co-founder of Carbon Collective, a startup financial advisory service focused on solving climate change, to give Jasper a whirl at writing the specialized copy required for his site’s blog. It was a no-go. “Jasper’s output is too general and repetitive for our readers, who are looking for a trustworthy, deep dive into our subject matter,” Stein told me after fiddling with it for an afternoon.

My guess is there is not enough targeted content on investing in climate change for Jasper to study. 

Jasper’s job is a multi-step process that gathers, stores, cleans, mines, and prepares data for analysis. However, data that would take a team of humans years to collect, process, and prepare for use can be completely readied by data analytics technology in fractions of a second.

In order for it to work in a copywriting context, you must set parameters that help the technology zero in on your content goals by stating not just the what, but the why and how of your copy.

Though each platform is based on the same GPT-3 technology, the methods of generating copy vary among platforms.

When using an application like Jasper for copywriting, you would:

  1. Select the type of copy you’re after, such as headlines, email, product descriptions, or blog posts.
  2. Provide the application with specifics, such as product details, relevant aspects of audience persona, or market segment.
  3. Add information about sentence length, readability, and tone.

The more information you provide, the more detailed and contextual the result.

According to the company founders, Jasper has already read 10 percent of the internet and incorporated the characteristics of language, nuance and all, to accurately replicate human speech in more than 25 languages. The results are not only mindboggling, but it’s downright spooky to think where developers will take this technology next.

Use of AI-Driven Technologies for Copy Creation

There have been concerns raised about the security of this technology and how using AI to generate content will affect academia. Another worry is the effect of computer-generated writing on intellectual property and copyright protections. Generally, these same protections exist as soon as unique content is written and published, no matter who or what creates the content. However, copyright law may need to adapt to new creation concepts as technology evolves.

As far as professional copywriting is concerned, this should be considered more as a tool than a replacement for marketing copywriters. Copy bot-generated content tends to be wordy and often phrase things in a way that actual humans – or at least good writers – wouldn’t. Then there’s punctuation and a startling lack of variety in sentence length. Don’t get me started on that.

However, they’re ideal for brainstorming and creating marketing material that doesn’t change much, such as product descriptions.

Other good case use examples include:

  • Headers
  • Email
  • Social media posts

Jasper was initially trained on 200 billion words. Another GPT-3 based copy generator, CopyAI, uses 175 billion parameters to create marketing copy. The self-learning nature of the beast allows it to develop further to incorporate more nuances of human language, including context and grammar in multiple languages.

What that means is that the quality of output is almost certain to improve with time.

Despite the potential and increasing use, there are still some limitations.

  • Garbage in/garbage out: Because this technology can use the entirety of the internet for learning, it’s able to access the best. However, it’s still exposed to the worst of written communication that the internet is notorious for supporting as well. With all of the misinformation and content creators who can’t distinguish “there” from “their” or “they’re,” that’s a little scary.
  • Copied content: Although content is still protected by copyright law, having thousands of brands using the same technology to create marketing content could still lead to having content that’s too similar – or worse, identical – to your competitors.
  • You’ll lose your voice: This is another area that threatens the uniqueness of your brand content. Although an AI-powered copywriter can be trained to create unlimited amounts of unique copy, it can never be trained to reflect your unique brand voice.
  • It isn’t suitable for omnichannel marketing: We all know that the route of the buyer’s journey and consumer pain points differ from customer to customer and from one marketing segment to another. It would take more work than it’s worth to tailor each bit of robo copy in a way that refines text for the same product so that it reaches different audience segments.

For example, you would place a product differently for a B2B customer than you would if you were marketing to a regular consumer. You wouldn’t highlight the features of your running shoes the same way to an athlete as you would to your next-door neighbor.

Who would benefit from such an innovation?

  • E-Tailers with limited or small catalogs or unchanging inventory/product lines
  • Sellers marketing non-technical products or products that don’t need to list specifications
  • Those who lack time, help, or basic writing skills
  • Marketers who have a large client base
  • Those who need to generate a steady stream of relatively generic content on a regular basis, such as tweets, captions, and social media posts

Is AI Putting Copywriters Out of Business?

After thinking about it a bit, maybe Jasper doesn’t pose such an existential threat to me after all. One of the biggest sources of angst among content marketers and website owners is coming up with ideas and generating steady, evergreen blog content, emails, and ad copy. Copywriting bots can keep your content calendar full for as long as you’d like. They also eliminate common problems like plagiarism and writer’s block, and you won’t have to spend frustrating hours on SEO and keyword research.

Having a technology like this as a tool will make it easier to create targeted marketing content on a consistent basis. In short, it does exactly what technology was meant to do: aid professionals by completing repetitive tasks so that they can focus on business growth and product development.

Using this tool, you’ll also cut down on the amount of research needed when you’re writing about an unfamiliar topic or marketing to an audience in an industry with which you have little experience or knowledge. Keyword research will also become a thing of the past because trending words and phrases are already baked into the system. This means you’ll have optimized, consistent, targeted copy ready when you need it, and with a fraction of the effort.

With all of that free time at my disposal, I might even be able to finish that sci-fi novel gathering dust in my desk drawer.

About the Author

Bernard Brode has spent a lifetime delving into the inner workings of cryptography and now explores the confluence of nanotechnology, AI/ML, and cybersecurity.

Sign up for the free insideAI News newsletter.

Join us on Twitter: @InsideBigData1 – https://twitter.com/InsideBigData1

Speak Your Mind

*