The concerns recently voiced by marketing professionals about Large Language Models (LLMs), like ChatGPT and Google Bard, echo those of translators in the mid-2000s when machine translation tools like Google Translate first emerged. Translators were extremely worried that these sophisticated tools would render them obsolete. Yet even though the effects have been substantial, twenty years later the translation profession is still alive and kicking.
ChatGPT offers extraordinary capabilities that are set to disrupt the marketing industry in a similar way, causing a great deal of disquiet among marketing professionals. But can marketers look to the translation industry for reassurance? Are there lessons to be learned from translators who have understood and assimilated new technology?
To answer these questions, we’ll examine how ChatGPT is replacing certain human marketing activities and why this may put jobs at risk. Then we’ll explore some of the tool’s limitations and how they are similar to those encountered with Google Translate. We’ll then review how translators have adapted to the technology to demonstrate why the future for marketing professionals may not be as bleak as some think.
Marketing’s brave new world
Many marketers are already leveraging the power of ChatGPT to boost marketing efforts. A few simple prompts can produce impressive results that have the potential to radically change how marketing departments operate. As well as providing chatbot capabilities to improve customer service, ChatGPT can carry out a number of other key marketing functions.
Data-driven campaign strategies
The tool can be used to devise marketing strategies by entering specific campaign goals and commanding the tool to collect and analyse vast amounts of data on competitors, customer behaviours and market trends. If prompted correctly, the tool will then come up with a range of strategies and provide detailed explanations on how to implement each measure.
ChatGPT can be used to run A/B tests by generating different versions of marketing messages. It can then analyse the responses from customers to optimise campaigns and identify the most effective variations.
The AI tool can quickly analyse huge amounts of customer data and make personalised content recommendations and messaging to enhance customer engagement and conversion rates.
ChatGPT’s natural language processing capabilities enable it to generate coherent content, such as blog posts, social media captions and email newsletters. It is even possible to prompt the tool to write in a particular way, such as “in the style of copywriting legend David Olgivy” or even Shakespeare himself if it fits the brief.
With all these impressive capabilities, it is easy to see why some marketing professionals feel they are facing an uncertain future. This was exactly how translators felt when advances in machine translation accelerated rapidly in the mid-2000s. Back then, experts were predicting the demise of the profession, but this never happened – it merely changed. This is likely to be the case for marketing and other professions such as software programming and teaching that are expecting significant disruption due to AI technology.
It is very important to look beyond the hype and scaremongering. ChatGPT’s limitations are very similar to those that quickly became apparent after the launch of machine translation tools like Google Translate. Translators soon realised that the tools could not outperform them in a number of aspects and would be unlikely to do so for some time to come. Surviving as a translator has been based on understanding how to use the technology and guard against its limitations.
The hallucination effect
The developers of the most advanced machine translation tools such as Google Translate and DeepL advise against publishing machine-translated texts before they have been checked by a professional translator who has been trained to detect errors. Most errors are caused because the tools are incapable of understanding context and ambiguity. In a similar vein, it would be very unwise to base important marketing decisions on ChatGPT’s output or to publish marketing content that has not been fact-checked and edited by a professional marketer.
The developer of ChatGPT, OpenAI, has acknowledged that “ChatGPT occasionally produces answers that sound plausible but are incorrect or nonsensical,” a statement that has been confirmed by numerous users. This is because not all of ChatGPT's training data has been verified. Furthermore, if ChatGPT does not know the answer to a prompt, it tends to make something up.
It is easy to be fooled because the output sounds so authoritative. This is sometimes referred to as the ‘hallucination effect’, where AI systems generate outputs that appear plausible but are not accurate or based on real information. This is exactly what happens when using machine translation tools. The generated translation looks totally convincing because it reads in a very natural way, however, it is often not an accurate rendering of the original language.
In the same way that experienced translators never trust machine-translated output, seasoned marketers are unlikely to ever make important marketing decisions or publish material based on ChatGPT’s generated responses, however convincing they sound.
Security and confidentiality
Experienced translators are always mindful of the security and confidentiality issues associated with online machine translation tools such as Google Translate. This is especially the case when handling client data in the medical, financial and legal fields where the stakes are high. To avoid these risks, trained translators use offline tools or paid cloud services where there are stricter measures in place around security and confidentiality.
Similar measures are likely to be applied by savvy marketers when using AI, especially now that cracks in ChatGPT’s security have been exposed. In March 2023, the Italian data security authority banned its use after user prompts had been hacked. It is very unlikely that companies will be willing to upload sensitive marketing intelligence for processing on a platform that could be easily breached.
The human touch
Machine translation tools often struggle to reproduce the required tone of the original content, so human translators are needed to inject the human touch, especially when it comes to marketing content. Machine translations often lack the style and personality required to evoke the desired emotional response from the audience.
The same principle also applies to marketing content that is produced by Large Language Models. A blog post created by ChatGPT may include all the right information and be tailored to the target audience, but it may sound flat and uninspiring. It might also be out of step with an organisation’s other marketing materials. This means human marketing expertise will still be required to ensure the tools can successfully replicate brand voice in a consistent way.
It is also worth considering that there may be a risk that AI-generated content will be penalised by search engines. Google’s algorithms are designed to rank high-quality and relevant content more highly. Due to the variety of concerns surrounding AI, Google may update its algorithms to ensure human-generated content is prioritised.
Usurper or assistant? Balancing AI brilliance with human intelligence
Many translators nowadays work with machine translation tools such as Google Translate to help them improve productivity, check terminology and support quality control. However, this does not mean that we completely rely on them. They are just one of the many tools we use to do our jobs. We also draw upon translation memories, term bases, dictionaries, reference works, corpora, style guides, glossaries and parallel texts.
Similarly, marketers will not rely on ChatGPT to complete all marketing tasks. It will sit alongside CRMs, CMSs and a raft of other tools used to analyse data, create content and manage projects. The marketer will remain the spider at the centre of the marketing web. The strength of LLMs like ChatGPT lies in their ability to synthesise and reformulate vast amounts of online information, which can significantly reduce effort and thinking time. Therefore, instead of seeing them as usurpers, it would be more appropriate to view them as clever assistants that help to sort and formulate information.
Google Translate and ChatGPT are ultimately only as ‘intelligent’ as the people using them. An inexperienced translator will not identify errors in a translation in the same way an inexperienced marketer will not be able to spot statistical errors or branding issues. This means there will always be a demand for well-trained and dedicated professionals to check and refine output no matter how impressive and powerful AI tools become.
One door closes and another opens
Tools like Google Translate have radically changed the type and volume of work available to translators, who now find themselves proofreading and editing more often than translating. However, this has been partially compensated for by the exponential increase in online content that needs to be translated, meaning that good translators will still be in demand. New opportunities are also opening up that are allowing translators to diversify and acquire other skills in areas such as content writing and editing, SEO translation, software localisation, subtitling and cultural consultancy.
Similarly, it is likely that even though certain tasks will become obsolete or change for marketing professionals, new opportunities will emerge in areas such as ethical compliance and AI strategy.
A different outlook
We hope that examining the similarities between the limitations of ChatGPT and Google Translate may provide some reassurance to marketing professionals who are fearful of the impact of AI tools such as ChatGPT. Not only does machine-generated content tend to lack the nuance and emotional impact that humans can bring to the table, but it also poses significant risks in terms of accuracy, security and confidentiality. Human oversight will be needed at all times.
While the future of marketing will undoubtedly be significantly influenced by AI tools like ChatGPT, it is the human touch, expertise and adaptability that will continue to be valuable assets. By embracing the strengths of AI and combining them with human intelligence, just like translators, marketers can navigate the evolving landscape and seize new opportunities to safeguard their futures.
This blog post has not been generated by ChatGPT 😉