Much ado about AI content generation

an image by Tara Winstead that shows a human hand reaching out to a robot's hand

I know I said I don’t have a promised cadence for my writing. I’m not sure what to call them. But I’ve set up a time in my calendar to write between 6–7 pm on Sundays and titled it “Review your week in public.” Thinking of these as reviews makes it less intimidating.

So here goes:

Something I read that warmed my heart:

  • If Only You by Chloe Liese

Something I listened to that made me dance:

What I’ve been thinking about: Artificial Intelligence (AI)

I haven’t used ChatGPT yet. I plan to test it out sometime soon, though. As a writer, marketer, and student, I haven’t escaped the doomsday proclamations about AI. Combined with the tech industry layoffs, I’ve often wondered if I’m being too slow to realize that AIs will take over my job soon, and maybe I should have listened to my parents and become a medical doctor.

I must admit that I am both skeptical of AI’s impact and pessimistic about what it means for the future. I think it will enable bad writing/marketing practices and fuel the overproduction of low-quality content we already suffer from.

Think about it, Google’s mission has been to make the web more accessible and useful since its creation. Yet we still have a ton of content online, on Page 1, that just makes me want to tear out my hair.

The other day, my headset broke in a freak accident, and to distract me, my partner wanted us to play the 21 questions game virtually. Since neither of us was exactly sure how the game worked, we tried to read articles that could show us sample questions, and nothing we found made sense.

One of the articles told us that in a game of 21 questions, all the responses had to be yes or no. Tell me why one of the examples was “Would you rather x or y?” How is that a yes or no question?

It also reminded me “unfondly” of my content mill days when the client would ask me to write the same article in 10 different ways. I ended my contract with that client when I was broke and never regretted it.

Why I’m skeptical of AI

One of my projects in April was to repurpose a webinar into an article. An AI cannot turn that webinar into a narrative-driven article that aligns with the product’s messaging the way I did.

  • First, I had to create a writing voice for the piece. Some things the speakers spoke about would not translate verbatim in writing, or it would read like they were rambling.
  • I also restructured the content to align with our content style and have an interesting flow for a reader rather than a listener.
  • I decided to cut out some areas that were not relevant to the piece, match them with old quotes from the speaker and also infuse our product at the right opportunities.

AI lacks creativity and can’t make some human-centered decisions. It doesn’t have context for my writing, and since I don’t aim to produce another piece of content that doesn’t make a difference to the internet, our target audience, or the company’s bottom line, I’ll pass.

When I’m not repurposing a webinar, I’m talking to internal and external subject matter experts, reading and critiquing primary research, synthesizing the data and insights I’ve gotten, and reviewing my work after it has been edited to ensure that it’s valuable to prospects and the sales team can point to the article as something that helped close a sale.

If ChatGPT can do that, show me.

The ethical implications

AI tools are not intelligent by themselves. Since it lacks the innate creativity that comes with being human, it has been trained based on the writing, conversations, books, films, and art humans have produced over the years. None of these people reap the benefits of their work being used to train these tools. We now have artists, authors, and record labels scrambling for policies to protect their work.

It has also not helped that the conversation on AI has been very profit-driven. I’ve seen several posts on LinkedIn and Twitter from business owners who think they don’t need to hire writers anymore. This undertone of glee at layoffs just rubs me the wrong way.

It’s an employer’s market, power has swung back to them post-Covid and the quiet quitting trend, and they can now point to AI as the driver for their business decision’. I don’t believe the claims that AI is now capable of replacing the millions of workers that have been laid off. I predict there will soon be an onslaught of pieces on burnout in the workplace again, with one person having to cover the labor gap.

There are even more concerns about deep-fake technology and how it’s being weaponized more and more against women. For example, where a nude photo was generated and shared online with a real person’s face. Some people pointed out it was fake. Some claimed it was real, and others didn’t care.

Some AI content detectors have attributed human-written content to AI. So writers now have to defend their work to clients who believe the AI detector instead. There was also the case of the audio that circulated of a presidential aspirant in Nigeria begging a megachurch pastor for support. One of the reliable media houses fact-checked it to be real. However, they used the wrong tool to detect deep-fake technology, making their claim unreliable. These examples reveal a considerable gap in educating the public on the limitations of AI and the need for more control over how it will be used.

My friend, B, argues that AI can give Africans an edge in the marketplace, and we should focus on that rather than ethics. I disagree. The tool can only augment what we already know.

My friend is brilliant and has a keen editor’s eye; AI will be more useful to him than someone who isn’t a skilled writer. He can spot the gaps in what ChatGPT feeds him, rewrite sentences to fit into context, and fine-tune prompts to get what he wants.

An average or newbie writer will not have the same outcome. And resorting to tools might only impede their skill development.

As I see it, AI might take jobs from African talents and increase the entry barrier. The people who would rather use AI to generate 20 articles in one hour are not very different from those who only hire African writers when they want to pay criminally low prices.

On the one hand, we can say that what you get on content mills matches the low pay. But I can tell you from experience that those clients still expect a certain quality. And the ones most common in mills are similar to the ones AI has produced. So it will only reinforce using AI instead of hiring African talent.

On the other hand, even when you produce quality work, some clients will reject you or want to pay you peanuts because you’re African. AI can’t control other people’s biases. In fact, case studies show that we’ve transferred our biases to it. So I think we need to look at or develop more use cases on how AI will help African talent.

I like the idea of AI as Augmented Intelligence, playing the same role technology has occupied over time, i.e., enabling and amplifying human creativity. In his piece for the New Yorker, Lanier presents a humanistic view of AI that I agree with and feel more hopeful about —

We can now imagine a Web site that reformulates itself on the fly for someone who is color-blind, say, or a site that tailors itself to someone’s particular cognitive abilities and styles.

He dived deeper into ethical concerns and possible policy development on AI; please read his piece for more.

As the tool advances, I’m sure there will be improvements in what it can do, and I’ll be looking out for them. I’m thinking of writing a research paper on how it can improve marketing or help media houses in Nigeria. I’ll publish it here if I get around to it.

Recommended reading

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.