AI and the future of content: 5 key considerations

However, far from being a passing novelty itself, AI is set to fundamentally alter content production. We carried out extensive research into the use of AI for global content production to offer some insight into the accuracy of ChatGPT-4 and attitudes toward AI.

Although the focus of our editorial testing in our report was its accuracy in localizing content versus human translators, we’ve also tested across a range of long-form editorial creation, too – and so will cover a few general thoughts on creating content with ChatGPT-4.

What is high-quality content?

Our content offering is all about organic performance – and as such we need to consider Google’s approach to high-quality content. To rank well, you need to consider their E-E-A-T criteria: Experience, Expertise, Authority and Trust. The best content, which is in the top positions on the search engine results page, is accurate, interesting, and generates multiple backlinks.

AI versus human translation

To test both human and AI localization thoroughly, we created a sample article about UK motorcycling culture. A topic with many tricky local terms, slang phrases and very specific references that can be tricky to localize.

We then tested localizing the content into French, Spanish, and Simplified Chinese. Human translators and ChatGPT translations were evaluated based on accuracy, language and consistency, with error types broken down from critical to preferential.

Human translators outperformed ChatGPT in all three languages tested. ChatGPT found Chinese the most difficult, with a 28% drop in accuracy compared to human translation. The Chinese content translated by AI included significant mistranslations and broken phrases.

AI-generated content and trust

We’re in the business of marketing. Any activity we embark on should have high confidence of succeeding. Yes, risks are good sometimes, but the key is to not alienate audiences. Trust is a big factor here.

Our survey, which focused on consumer trust in AI-generated content, involved 2,500 participants from the UK, the US, France, Spain, and China.

A large majority of people (73%) felt positively or very positively about AI tools in general. The positive sentiment correlated strongly with consumer awareness of AI and was highest in China, at 94%.

Trust in AI-generated content varied across regions and industry types, but the biggest issue for consumers was transparency. While trust in brands that use AI remained strong, 80% of respondents believed that brands should disclose their use of AI.

Here are five key considerations for using ChatGPT-4 for content.

1. SEO: what would Google think?

Google is pushing hard for authorship to play a big role – even bringing author bios into the results page. Coupled with the best practice recommendations to disclose AI usage, it seems to point to a clear separation of what’s AI and what’s not. Both scenarios could represent a blend, but Google clearly views quality content associated with actual human authors as a signal of trust.

2. An efficient toolkit

ChatGPT is especially useful for analyzing data sets such as keyword research or helping understand the organic opportunity in specific markets. The caveat is that it does not have access to the latest keyword volumes as it operates based on a static snapshot of the web from 2022.

3. Cultural nuance

AI language models ‘learn’ from existing content and produce text based on a prediction of what word is most likely to appear next. The nature of this process lacks cultural nuance. At best, this leads to ‘clunky’ copy; at worst it can lead to consumer offense where cultural, religious and societal norms are not understood. In a market where consumer trust is essential, local knowledge remains vital.

4. Brand training

AI-generated content can feel wooden and lack dynamism. In a world where the tone of voice is crucial, ChatGPT and other AI tools need ‘training’ to deliver content that represents your brand well. Skills in crafting prompts and helping AI to learn your tone of voice will be essential to harnessing the potential of AI for global content production.

5. Clear frameworks

Consumer receptiveness to AI-generated content is high, but so is a desire for transparency. Brands should think holistically about their use of AI and, with policies and processes in place, can talk confidently about how they use AI to serve the consumer more effectively.

And finally, a healthy dose of skepticism

Never forget that the unique perspectives and ideas that will help your content rank on the first page and generate backlinks for your client can only truly come from an actual expert in the field. An expert that will help the all-important authorship strategy and Google’s E-E-A-T criteria.