Rise of the machines: AI in marketing
The genie is out of the bottle and there’s no going back now – the rise of the machines is upon us! Since November 2019, AI has been all anyone in marketing can talk about. ChatGPT dropped into our lives and many marketers jumped on it, using it for data analysis, content generation, ideation and more. But this has raised debate around the ethics and efficacy of AI within marketing, and beyond.
AI has rapidly evolved from niche to mass-market with adoption jumping to 72% of companies across various sectors.[1]With this popularity comes the acknowledgement that, although a valuable tool, it is undeniably flawed. So, how can it be used for growth, where do we draw the line, and will this lead to an all-out war of humanity against technology?
Terminator or tool?
So where to begin? AI is split into two distinct groups: analytical and generative.
Take toy company Mattel, who use generative AI for ideas in their production of Hot Wheels, to inspire new toy car features and designs.
Or cereal giant Kellogg’s, who use analytical AI to scan trending recipes online that include, or could include, cereal, to create social campaigns. And we all know Netflix analyses what you watch so it can offer up personalised recommendations of further viewing, attempting to calculate which genres, actors and writers you prefer (not always on the money… that’s AI for you!)
Many decision-makers are looking to both tools to economise budgets and reduce spend. Current gen AI has the potential to automate 60 to 70% of time-consuming tasks[2], including personalisation of messaging and analysing and predicting customer behaviour. But this naturally leads to some employees worrying about its implementation, as they fear it could be coming for their jobs, rather than allowing their creative functions to be pushed to the fore.
The real opportunity is in growth, experimentation and innovating new ways to add value. About 75% of the value that generative AI could deliver falls across four key areas: customer operations, marketing and sales, software engineering, and R&D.[3]
Pros and Cons of AI Implementation
65% of organisations are now using generative AI in at least one business function (up from 1/3 last year). In marketing it can be used to automate competitive research, analyse website traffic, interpret customer behaviour patterns, and predict the performance of campaigns. This can free up marketers’ time to do those things that require a more human touch, like refining user journeys and campaign tactics to improve conversion rates.
It can also personalise customer experiences by analysing data and making recommendations as well as performing predictive modelling (i.e. “last month you ordered ‘Eat Healthier in 30 Days’, would you like more healthy eating recipes this month or can we tempt you with desserts?)
Another area it can assist in is saving on content generation costs. Not by eliminating the need for designers and writers but by assisting in content ideation and first draft ideas. Using AI as a starting point for a structure, or to see what information it collates, can be a time saving exercise that aids in the development of a piece of writing. Around 82% of marketers say that using AI for brainstorming content ideas has saved them at least one hour of their time.[4]
Of course, this doesn’t mean AI is the silver bullet to all your problems. It still has its own pitfalls and can’t be relied upon to do its thing without any human oversight.
The battle of bias (and inaccuracy)
Gen AI models rely on large amounts of data, so making sure you have data security policies, checks and balances in place is of vital importance to make sure both yours and your customers’ data is safe and secure.
As businesses increasingly implement AI into their day-to-day practices, they must also recognise the risks of its use. Inaccuracy and intellectual property infringement are core concerns due to its nature of gathering and repurposing content, and this also leads to a problem with biases in the algorithm.
Research conducted at The Hague University of Applied Sciences looked at the use of generative AI in creative design processes and found that AI could be open to cognitive, framing and prompt biases.[5]
The internet is a wild West of “interesting opinions” and “impassioned discourse” and AI may draw on these views to generate problematic or insensitive outputs. Mattel’s campaign to promote the Barbie movie saw gen AI Midjourney used to generate imagery of Barbie dolls from around the world. This received backlash when it generated a doll from South Sudan adorned with weaponry,[6] distilling complex geo-political issues into an offensive stereotype. Even ChatGPT openly recognises that its models can reflect social biases and prejudices.
For example, after three years of collaborating with IBM, McDonald’s introduced AI to take drive-thru orders at over 100 restaurants across the US. However, the AI would often get orders wrong and, in one case, kept adding Chicken McNuggets to an order until there were 260 on the ticket (that’s a lot of sides!). The company shut down the AI ordering system in June 2024 but still sees a future in it as a voice-ordering solution.
Beyond this, sometimes AI just gets it plain wrong. Without the human ability to pick up on context clues and guide it, it can run into some tricky situations.
Humans, of course, have in-built biases too, and make mistakes, but if you’re using AI, you need safeguards – and real people – in place to make sure those biases aren’t seeping through.
A creative team can correctly identify nuance and can understand briefs with subtlety that AI may miss. When it comes to creative campaigns, our motto at The Creative Consultancy is “we do human” … because B2B businesspeople are still people. They eat fast food. They find things funny. They’re tired on Monday mornings. And tapping into that humanity, warmth and emotion is how you get serious marketing cut-through, something AI still can’t replicate.
Contact us today and let’s discuss how we can create something bold and memorable that delivers real results.
[1] Alex Singla, Alexander Sukharevsky, Lareina Yee, Michael Chui, Bryce Hall, McKinsey, https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai, 2024
[2] Michael Chui, Eric Hazan, Roger Roberts, Alex Singla, Kate Smaje, Alex Sukharevsky, Lareina Yee, Rodney Zemmel, McKinsey, https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier#introductionm 2023
[3] Michael Chui, Eric Hazan, Roger Roberts, Alex Singla, Kate Smaje, Alex Sukharevsky, Lareina Yee, Rodney Zemmel, McKinsey, https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier#introductionm 2023
[4] Hubspot, https://offers.hubspot.com/thank-you/ai-marketing?hubs_signup-url=offers.hubspot.com/ai-marketing&hubs_signup-cta=Submit&hubs_offer=offers.hubspot.com/ai-marketing, 2024
[5] Andreea-Roxana Popescu, Alice Schut, Design Research Society, https://dl.designresearchsociety.org/cgi/viewcontent.cgi?article=1326&context=iasdr, 2023
[6] Sabrina Lynch, The Drum, https://www.thedrum.com/opinion/2023/08/30/ai-gave-sudanese-barbie-gun-beware-the-bias, 2023