Interviews, insight & analysis on digital media & marketing

The limitations of AI in marketing

by Sam Martin-Ross, UK Managing Director of digital marketing agency, Eskimoz

Artificial intelligence has gained the reputation of the new shiny toy in marketing, and as such, there’s an increasing tendency towards presenting AI as the silver bullet solution to all our marketing woes.

There are undoubtedly areas of marketing that AI excels at, far surpassing what human marketers are capable of, like parsing vast amounts of data, automating and optimising campaigns, not to mention its ability to revolutionise paid media campaigns through superior prediction and bidding capabilities.

However, the misapplication of AI tools and technology can pose a serious problem for the development of long-term and sustainable marketing strategies.

Content concerns

The rise of AI platforms like ChatGPT has seen a huge swell in awareness around the content capabilities of these tools. Many have seen this as a potential tactic to fast-track their content strategies and climb the search engine rankings, but in reality, this simply isn’t the case.

Using AI in SEO blog posts or website content in an attempt to rapidly rank in the SERPs may certainly have an initial impact, but in the long-term, Google quickly catches on to what you’re doing.

SEO is a cornerstone of marketing and requires continuous work and input to ensure you defend your position in the rankings. While it may be tempting to optimise this through AI-generated content, taking this blanket approach to the mass-production of content is flawed. It, of course, lacks the creativity and ability to contextualise copy that a human writer possesses, but deeper than this, it can actually harm your brand.

It has the potential to create biased, offensive or factually inaccurate content, damaging your brand reputation in the process, not to mention the impact of poor-quality content on your website’s conversion rates.

Data constraints

While AI is capable of doing incredible things for marketers, it is only as good as the data it is trained on. One of the concerns here is that it is entirely possible for AI tools to operate based on very small amounts of data, which in turn makes any conclusions drawn from these data sets risky at best and wildly inaccurate at worst.

This means that businesses, particularly smaller ones, that lack rich data sets could be basing critical decisions about marketing and long-term strategy on unreliable data.

More generally, because AI has this reliance on the data it has already been fed, it makes it very difficult to utilise these tools in response to sudden or ongoing consumer or market changes, as it lacks the reactive adaptability of human marketers.

AI may still have the edge over us when it comes to automating and optimising campaigns once set in motion, but the content and direction of those campaigns rely on human oversight and expertise to apply those data-driven insights.

Cost limitations

It’s hard to deny that many companies are already doing truly amazing things with AI and automation technologies. Nvidia, for example, has leveraged this technology to create 3D scans of products in their advertising, synthesised by AI through inputting detailed brand guidelines.

These kinds of applications have the power to reinvent how we market, but currently, for many businesses, advertising at this scale with AI technologies is prohibitively expensive, and for many, a less effective use of budget than many traditional marketing activities.

Compounding the cost issue, businesses must not only make the initial investment into AI tools, their operation and maintenance, but also ensure they have the right skills in-house to fully realise the potential of these tools. With marketing budget stretched thinner each year, this kind of investment may be too significant for many firms to justify.

Balancing SEO and AI

Google has also made some huge leaps when it comes to embracing the potential of AI technologies and integrating them more closely with its platform. However, this has thrown up some significant concerns from an SEO perspective.

Take AI-generated snippets, which use AI tools to improve search results and display information on the search page. These are great news for users wanting a rapid answer to a quick search or query but are not so great for the websites who are providing that information but being denied those all important clicks and boosts in traffic.

Google’s business model still relies on pay-per-click advertising, so it’s currently unclear how they will balance increased AI capabilities that enable content from websites to be displayed on the search results page through AI tools, with the need for websites to receive their fair share of clicks and traffic. As these technologies mature, they will undoubtedly shape new approaches to SEO.

AI is already having a huge impact on the marketing space, but it is not without its limitations. The danger lies in regarding AI as the answer to everything, rather than one more useful tool in the marketing arsenal. Looking at where humans and AI can work together, rather than taking shortcuts that bypass human creativity is the key to implementing AI tools in the long-term.