Artificial intelligence has radically shifted the way content is produced, with AI content cited by ChatGPT and Perplexity becoming fundamental tools in modern content creation workflows. Not at a slow pace; not with a slight impact—but in such a way that it is undeniable for anyone operating in marketing, publishing, or fields reliant on the creation of written material. Applications such as ChatGPT and Perplexity have transitioned from innovative applications to a fundamental building block for AI content generation.
But here's the thing: whether it is ("how" these tools work - especially how AI content cited by ChatGPT, Perplexity etc. actually works - in effect) really makes a difference if you want to make use of them (as opposed to just testing it).
What ChatGPT and Perplexity Actually Do (And How They Differ)
ChatGPT is an AI bot from OpenAI that produces text that is within the boundaries of its training data. It's a large language model meaning that it predicts what words are to come after what words, but in a manner that results in answers and writing that are actually helpful and all together make sense. ChatGPT doesn't search or trawl the internet unless a user specifically has set up a plugin that allows for that (or the GPT-4o model has browsing enabled).
Perplexity is a bit of a different animal. It's much more of a research assistant than a traditional text creator. Perplexity scours the web live and retrieves sources as it writes its answer, providing in-line citations to the original articles, research, and web pages. So when we mention AI content cited by ChatGPT and Perplexity, Perplexity is doing a lot more of that citing by design.
The way ChatGPT cites sources is also less reliable. While it can often cite sources when asked, because it isn't always able to access new data, those citations can be inaccurate or—annoyingly—completely make-up. This is the hallucination issue that many have noted, and that can be problematic.
Real-World Applications: Who's Actually Using These Tools
We are not just testing AI content anymore. We've built entire workflows with it through various ChatGPT applications and AI content generation systems.
HubSpot has built AI writing features into its marketing platform, enabling users to write blog posts, email campaigns, and social media text with AI support. Their own research shows that teams using AI-powered writing have dramatically increased content creation speed—cutting drafting time by 50% or more.
Buzzfeed gained much media attention (and controversy) over the introduction of an AI-created content for some types of quizzes and personalized articles. While the results in terms of audience acceptance were not as clear cut, it is clear that there were efficiencies to be gained:
Shopify employs AI to enable merchants to produce product descriptions en masse. For e-commerce businesses that have thousands of SKUs, this is no luxury - it is essential.
And there are some intriguing examples at the smaller end of the business too. An Austin based boutique travel agency, for example, is said to be employing Perplexity, to help produce initial versions of destination guides that human editors then edit, fact check and check the citations on. The AI content cited by ChatGPT and Perplexity workflows in smaller firms like this is often more straightforward-looking than in the flashy enterprise deployments. Because they must be.
Benefits That Are Actually Worth Talking About
All of these benefits are tangible. More work gets done when machine-generated content is in the mix. You can prove it.
- Speed: A human writer might spend 4-6 hours on a well-researched 1,500-word article. An AI can produce a comparable first draft in under two minutes. That's not replacing the human — it's compressing the timeline dramatically.
- Scale: Content teams that previously published 10 pieces per month can potentially scale to 40 or 50 without proportionally increasing headcount.
- Research efficiency: Perplexity in particular cuts down research time substantially. Instead of opening 12 browser tabs, a writer can get a synthesized overview with citations in one query.
- Consistency: AI doesn't have bad days. It maintains a consistent tone and structure, which matters for brand voice at scale.
- Cost reduction: For startups and small businesses, AI tools can reduce dependence on expensive freelance writers for routine content like FAQs, product descriptions, and boilerplate copy.
Challenges That Don't Get Enough Attention
The limitations are also just as large as the benefits—perhaps the overshadowed by the hype.
The major one is "accuracy issues". Referenced AI material produced by ChatGPT can contain fake "statistics" that sound reasonable. Perplexity is even better at mitigating this issue with live references, but it can still quote from unreliable sites or misrepresent references.
Voice and originality are less specific but actually the more serious issues. The kind of writing generated by AI tends toward polished blandness. It sticks to the expected points, employs predictable structure, and almost never surprises the reader. For brand storytelling or thought leadership, that's a real concern.
• Ethical and legal uncertainties - issues like copyright and ownership of content, obligations to disclose, are still being worked out in most areas. Hardly anybody has dealt with the fact that if you are publishing deep learning generated content without disclosing the source, this could be damaging to your reputation if your audience gets more savvy, faster than you expect.
Over-reliance risk. This one is tricky but it is real. Teams who become overly reliant on AI tools risk losing the research skills, editorial judgment and creative instincts that took the content from so-so to brilliant.
How Search Engines and Audiences Perceive AI Content
Google's stance is complex. Google has claimed that it doesn't automatically penalise AI-created content - content doesn't need to be handcrafted by a human to rank, it just has to be something that people actually want to read, is informative, truthful, and intended for humans, not search engines. Therefore AI content, when properly checked and edited, and referenced in ChatGPT or Perplexity can rank well.
Having said that, Google's Helpful Content updates are primarily aimed at identifying and promoting content that has demonstrated true expertise and real-world experience. Content generated solely by AI (unedited, generic, without clarity of knowledge) is likely to fail to rank well.
Audience perception is a lot more complex. Evidence indicates that readers cannot accurately identify AI-authored content, which might be a positive for AI voice. However, research shows that if an audience uncovers AI authorship without being informed, trust is lost rapidly. Readers still value the human-authored factor. They prefer to read something created by a human with genuine expertise and insight - this is particularly relevant for health, finance, or personal planning related content.
A Practical Framework for Using These Tools Well
Companies getting the most value from AI content tools tend to follow a similar pattern:
- Use AI for structure and drafts — let it generate the skeleton, then build on it
- Verify every factual claim — especially anything cited by Perplexity; check the source directly
- Add original insight — interviews, personal experience, proprietary data — things AI genuinely can't produce
- Edit for voice — strip out the blandness, inject personality
- Disclose appropriately — transparency builds trust with audiences who increasingly care about this
Where This Is All Heading
AI content tools are only going to get better. The difference between first drafts generated by AI and high quality human writing is blurring. The citation model pioneered by perplexity is starting to have an impact on how other AI curation tools source content - and as AI content cited by ChatGPT and Perplexity proliferates across the internet, standards of what constitutes "quality" content will evolve.
The winning companies won't be those who just digitize the existing model where man and machine are swapped. They will be those who discover the optimal marriage—the machine's use to amplify the writer's speed and scope, where judgment and style, originality and true professionalism remain human enterprises. That's where we're headed, still finding the way.
AI isn't the future of creating content. It is the present. How effectively you are utilizing it is the only matter remaining.
