Can AI Write Thought Leadership Content?
This article was originally published on March 12, 2023. It has been updated to address the evolution in capabilities of generative AI. “We’re Living Through the Boring Apocalypse,” reads a headline from a 2021 New York Times opinion piece by the famous psychologist Adam Grant. This article was in response to the repeated blasting of emergencies during COVID-19 and how that was making people pay less attention by the day. I’m starting to think that this applies to almost everything – especially B2B content marketing. The attention spans have dwindled, and the constant reiteration of uncharacteristic content justifies that. So, it makes sense why a lot of focus is lent to “thought leadership content” — perhaps the most credible, unique, innovative, insightful, expressive, and reliable source of information out there. Thought leadership content: However, this meaning has dwindled a bit in the past few years. The phrase “Thought Leadership” is thrown around so often in the social space today that it seems as if it’s losing its purpose – because people often don’t understand what it entails. Now that AI writers have graced the space, no wonder they’re being looked at as the means to amp up the thought leader content pipelines. Who knows if this space might become a commodity that can be completely outsourced to AI in the forthcoming months? But should we be doing that? This needs thinking and rethinking because the answer has to be sourced by dismantling the elements that lay the foundation for thought leadership. That’s what we’re going to do here. The Source of Content Where is the content coming from? A great thing about thought leadership is that it allows businesses to forge a relationship between their capability and the needs of the respective discipline they’re serving. This relationship depends on a host of things: These facets have one thing in common – they’re driven by human experience and insight. So, ultimately, the source of content is human. This can take the shape of: That’s just a brief list of options. The point is that thought leadership content is sourced from people for people. Now, let’s bring AI into the mix. So, even if you manage to write a blog using AI, you’ve probably edited it numerous times to make sure that it’s brand-specific, insightful, and relevant. But at what cost? Because it still wouldn’t be as good as something you’ve written from scratch. That brings us to our second facet — writing. The Narration By now, we understand that AI is not to be referred to for the information that’s going to be laid across the length of the article. But what about writing? Considering we have all the relevant information and a profound understanding of the direction we need to take, how about using AI by giving it prompts and providing cues to generate content at speed? Sounds reasonable, isn’t it? We experimented with this. But our research into the viability of AI writing tools for B2B tech content revealed that: The anecdote here? The AI writers aren’t making thought leadership content better – and they’re not easing the burden on writers, either. If anything, they’re diluting the narrative by making it less original. And there’s a big reason for the same. Thought leadership content isn’t actually a content type; it’s a practice — a form of narration that a business espouses to stand out and reflect upon its authority across a discipline. AI writers haven’t been programmed this way. Broadly, there are two categories of AI writers: What we experimented with were AI writing assistants – predominantly because we wanted more control over the narrative. The second category, i.e., long-form AI writers, are better suited for churning out extremely basic, SEO-focused content. The usefulness of such content is a debate for another time. All in all, across both categories, the problem of the sameness of content is profound – which goes absolutely against what thought leadership stands for – uniqueness and credibility. A Summary of the Discrepancies Hallucination AI can mix up facts, statistics, and even concepts. But it sounds lexically intelligent and convincing, so it’s challenging to discern whether something outlined is right. This requires subject matter expertise and constant monitoring. This problem, however, can be alleviated now with the use of ChatGPT plugins and web browsing capabilities. It will not help generate unique content, but it will at least provide a reference to the source of information. Deviation Thought leadership content in the B2B tech space, in particular, has a narrow and concrete focus. Because you’re not controlling the output beyond the prompt given or the manual editing after the AI has written, there are high chances that it constantly deviates from what you want to say. Recency When this article was originally written, it was challenging for AI to outline something that was recent. The dated data set didn’t serve the purpose in a space where narratives needed to be catered to the “present” concerns. However, this is not the case anymore. Generative AI tools, like ChatGPT, and AI search engines like Perplexity and Google’s SGE are now being powered by browsing capabilities. As such, they can weave in recent insights into their answers. Authenticity Well, carving out unique content is almost impossible because it’s all computation on AI’s part. The textual outputs are a result of excellent language correlation around a subject; however, there’s no personality or uniqueness. Data Storytelling The knowledge residing within your operational workflows and the successes that you’ve had with your clients is invaluable. It’s directly related to your business’s capabilities and how they serve the discipline in focus. AI cannot replicate this. Are There Any Benefits of AI for Thought Leadership Content? There surely are! It’s a balance of “productivity” and “originality” that needs to be sought here. Productivity can be amplified with the use of ChatGPT’s suggestions on a particular subject. Originality transpires from in-house insights and experiences. I often use