The 7 critical problems with AI-generated content

AI-generated content is seriously hurting brands

Artificial intelligence (AI) may be saving businesses time and powering game-changing, life-saving technologies. But AI-generated content isn’t doing brands any favours. To mark our 20 years in communications, PR and marketing, here’s why AI doesn’t always bring progress.

We’re big believers in technology. We work with some incredible businesses in the travel technology space that are innovating, redefining and leading hoteliers towards a brighter future. AI has been gradually making its way in, slowly at first, and now at lightning speed – just like we’re seeing in every other industry.

But there’s one aspect of AI that is hindering rather than helping progress. AI-generated content, created by the likes of ChatGPT, is everywhere. ChatGPT has 400 million weekly users, and OpenAI plans for that figure to be 1 billion monthly users by the end of 2025. It’s being used by people to reply to everyday emails, it’s being used to write entire books, and it’s being used for almost everything in between. Godaddy will produce entire websites with full content in minutes - this is the power of AI - and great for the self-employed and small businesses, but is it really good enough?

Putting concerns with quality aside for a moment, the environmental impact of this is a critical issue too, as every ChatGPT question uses 10 times more electricity than a traditional Google Search.

In our work to understand the marketing and PR opportunities for our clients, to examine the nuances of the landscape they’re operating in, and to craft content to push their brands further – we come across AI-generated content all the time. It’s everywhere. And it’s a problem. Here’s why.

1. AI-generated content isn’t original

Systems such as ChatGPT are trained on and powered by pre-existing, pre-published content. These systems cannot mirror human thought, make connections between things, and come up with brand new ideas. AI-generated content isn’t sparking the debate or furthering discussion. As time goes on, AI models are actually training on AI-generated content that’s been published on the internet, further increasing the potential for poor-quality content.

Case in point: If you ask it to summarise the key attributes of ecommerce, it’ll produce a near identical example over and over again. 

2. The content has zero personality

As much as you prompt an AI content generator to write in a particular tone of voice or to follow a brand style, the output never quite sounds human. This is exacerbated by the fact there’s no opinion, and that ultimately the output is a generic cross-section of information that’s already out there.

Case in point: The language and structure is formulated, continuously drawing on the same expressions. Just see this research done by our dear friends at Adido - How to spot AI generated content.

3. It might not be true

AI systems that generate content are known to hallucinate and create content that simply isn’t true. Although hallucination rates have gone down as the technology improves, hallucinations are unlikely to disappear completely. No business can afford to publish incorrect content, and you absolutely have to qualify the resources and sources AI-generated information is coming from – every single time.

Case in point: What are the sources of the content? What are the criteria you set for credible, reliable and trustworthy reference points? Make sure content stands the test of time.

4. Copyright is a murky area

The issue of copyright has been discussed ever since AI first started producing content, and things are changing all the time. At the moment, AI outputs can be protected as ‘computer-generated work’ under UK copyright law, and the person who organised the creation of the output is considered the author. But as the outputs are based on pre-existing content, plagiarism is a real concern. And no business or individual wants to be accused of copying or stealing content.

Case in point: AI content is based on content it has already digested. There are AI detection programmes which identify whether it’s original work.

5. Everyone can spot it a mile off

With AI at everyone’s fingertips these days, it's become very identifiable to many people when AI content is being used. There are also various AI content checkers out there that can flag AI-generated content. But many of us don’t need to use these checkers to notice whether AI has been used or not. Perhaps AI will actually force us to be more human and authentic. 

Case in point: Are you trying to build relationships, or create distance? Think about the impact it’ll have if you use AI when responding to emails, or providing feedback on someone's work. 

6. The quality can be terrible

Clearly the quality of AI-generated, written content can be very poor as it’s usually generic, unoriginal and lacking in personality. The same applies to AI-generated images. You see pixelated images, AI depictions of humans with blurry hands, and sections of images entirely missing when you take a closer look. Even on stock photo websites, AI images are often at the forefront. It’s misrepresenting the real world and removing us from reality. We’re a society that has such advanced multimedia capabilities – so why are we content to publish half-baked ideas and low-quality content? 

Case in point: Just look at these AI mistakes with written content, images and commentary on the rise of ‘AI slop’…and this is what we are beginning to see everyday. 

7. It doesn’t create impact

So we can produce content about anything in seconds. But are we taking the time to consider the purpose of the content? Are we defining what impact we’re striving for, and who with? AI is often being used to satisfy a checkbox. We need to be clever about how and why we're creating content. And this often comes back to crafting top quality content that serves a purpose and furthers a brand.

Case in point: One of our clients wanted to highlight key pricing features to improve usage of this functionality in a particular product. The brief was a generic advice article on ‘benefits of pricing’ without considering the audience – and you see AI-written content like this all the time. So we developed a more specific brief on ‘how tour operators can improve revenue through pricing’, which’ included well-tested use case examples from experts. We wrote the article – with no AI involved – and it went viral. 

We need to challenge ourselves

As our world has become filled with AI-generated content, our human mindset has become about quantity over quality. We need to expect better of ourselves. It’s one thing using AI for ideation; it’s another thing to actually publish it. How does that reflect your brand and its values? 

Returning to quality as the driving force of your communications, marketing and content strategy is what will enable you to stand out from the crowd and make an impact. Just because we can use AI doesn’t mean we always should.

With more than 20 years’ experience in Haynes MarComs, our clients have always valued our insights and the original, brand-building content we produce for them. Contact us to find out how we can help you stand out in a world of generic AI-produced content.