When Instagram garnered a million users in two and a half months – more than three times faster than Facebook – technology pundits were astonished. With the launch of ChatGPT, jaws hit the floor as the tool hit that mark in just five days. Its potential to save time, money and drive customer experience (CX) were obvious, and leaders across industries immediately refocused their attention to consider its business applications.
At the forefront were marketers, eager to explore integrations and put ChatGPT’s raw power to use. The benefits were immediate, but it was clear they had only scratched the surface. Yet, sometimes scratching an itch can make matters worse, and in the case of ChatGPT, the rush to put it into action could undermine a brand if it fails to keep tabs on the risks.
What can marketers do with ChatGPT?
ChatGPT can help marketers produce content at scale and make light work out of labor-intensive tasks. Its ability to analyze user data for personal messaging can create highly targeted campaigns and deliver equally high conversion rates. It has a role to play in customer experience too, including the use of chatbots to automate processes and answer queries, improving CX through faster response times and better support. AI-driven bots can even do things like “watch” Zoom meetings, pull salient points from the discussion and return a summary.
Still, at the top of the list of benefits is ChatGPT’s ability to create content for a wide range of marketing purposes. Blog posts, emails, whitepapers, product descriptions, advertising copy – the applications are extensive. If you hit a creative wall, you can use it to generate ideas. It can simplify research, boiling down the complex into easily understandable themes for consumers of information. It can also generate and assist in the management of areas such as social media, search engine optimization (SEO) and surveys.
So if ChatGPT can do all this, where do humans fit in?
What’s the risk of ChatGPT for marketers?
While ChatGPT is a game changer when it comes to gathering and distilling information, it’s not ready to produce content that’s slapped on a website or instantly distributed. ChatGPT cribs information from the Internet, which could be copying somebody else’s thoughts. So while it may save time and money, it exposes an organization to intellectual property (IP) and information accuracy issues.
When it comes to an industry that deals with sensitive information, ChatGPT can create serious problems. Recently, two attorneys wound up in Federal District court for filing a legal brief filled with fake judicial opinions and citations, courtesy of Chat GPT. In addition to being fined, the judge severely criticized the lawyers, ordering them to send a copy of his opinion to the very real judges whose names ended up in the fictitious filing. The fallout will undoubtedly end up costing them business.
Always keep in mind that ChatGPT’s AI is trained on the public Internet, which holds inaccurate, offensive and personal information. This can easily make its way into its output, subjecting a business and its clients to legal action and damaged reputations.
These risks are amplified by the challenges of policing its use. Leaders are putting their name on everything created by their teams, and younger generations are particularly eager to experiment with generative software. Though tools exist that can identify AI-generated language, they’re expensive, unreliable and can be time-consuming to use. Leaders need to make their company policies on generative language crystal clear, and ensure employees are only using it for safe, approved tasks.
Why does ChatGPT require human intervention?
ChatGPT is excellent for producing first drafts, whether it’s for a blog or an email campaign. Anything further is when humans need to step in to ensure accuracy, review for IP infringement and make edits. What’s more, ChatGPT can come up short on “feel.” Well-crafted material has a personal touch that automated language can’t yet replicate. For many, sacrificing that attribute isn’t worth it, no matter how much it reduces costs and effort.
And that’s if the content is even still relevant.
For the past two decades, Google has been the juggernaut for gathering information. In recent years, the company has made considerable updates to ensure it returns high-quality results – not duplicate material. The algorithms on search engines are able to incorporate new information much faster than generative data banks, which can result in ChatGPT delivering outdated or outright false output.
Take COVID-19 for example. Search engines were able to adapt to the influx of searches and quickly prioritize relevant, timely results, ensuring the latest developments were top of page. Generative models, however, take significantly longer to incorporate a critical mass of detail to create up-to-date language on the subject. This makes it difficult to count on these generative models for messaging that feels current. And at this point, Google still knows the difference between AI generated and authentic, fresh content. Companies using Chat GPT to write SEO content are putting themselves at risk of being punished by Google’s algorithm.
Human intervention remains critical with tools like Chat GPT, not just to ensure strong, relevant copy, but to oversee processes and make the right decisions on usage.
What is the future of ChatGPT and marketing?
It didn’t take long for the first major defamation challenge against ChatGPT and its parent company, OpenAI. A radio host from Georgia claimed the tool produced information that erroneously stated he had defrauded a nonprofit and embezzled funds. And since then, several high-profile authors including Sarah Silverman have sued for copyright infringement of their work. Though intent may prove difficult to assess, the ruling could have a major impact on the development and distribution of generative language tools, establishing precedent on liability, while raising ethical questions on intentionality and public impact.
The question of responsibility will affect how marketers leverage these tools in the future, and, how comfortable brands are putting their name on the content they create. But regardless of the outcome, the egg will not go back into the shell, nor do marketing leaders want it to. Generative language tools will make their lives easier and deliver newfound success, all while lowering overhead and speeding up production. Ultimately, clarification on responsibility will help to build proper guardrails and processes, not halt adoption overall.
Marketers are paying attention to the changes taking place at search engines, too. Titans like Bing and Google are working hard to integrate their generative language models and deliver better search results to consumers. This will transform how users find information, posing serious ramifications for SEO and SEM strategy and output. Firms that get ahead of that shift will be well positioned to take off and thrive, while those that are slow to adopt may get left behind.
Lastly, many risks of generative language tools like ChatGPT stem from their “black box” nature. Theirs is no way to trace the origin of information, making it difficult to ensure content is accurate and not confidential. Firms in all sorts of industries are looking to solve this through custom-built models with clear limits on learned materials. Tools like Perplexity and Waldo attempt to cite sources, offering at least some traceability. Other organizations are investing in proprietary tools that pull from a limited, siloed data pool. Though it may take time for innovations to reach the language quality of ChatGPT, marketers can better trust the content is not plagiarized or misleading.
Despite such legitimate risks, the future of ChatGPT in marketing is very bright. Marketers need to be exploring generative language tools today, as the benefits are enormous. At the same time, they must keep a close eye on usage and ensure human oversight in order to grow safely alongside the technology’s development.
