Edited by Peter Steidl, Behavior change expert

This article concludes a three-part series. You can find the first part here and the second part here.

In this third and final instalment I want to present my own view of the future, using my limited, non-artificial intelligence. There are several important aspects and developments that will define the future of marketing and advertising. These include an understanding of why –

The use of generative AI to assist creatives to develop concepts and shape ideas, create headlines and copy, design websites and social media posts and more has rapidly become mainstream practice. Naturally, the vendors of such applications and the creatives utilizing them emphasize the tremendous value they bring to the table.

However, there are limitations:

First, cognitive fixation has been identified in various studies and applications as limiting ideation. Cognitive fixation is a process where a stimulus, such as an idea offered by AI, takes the thinking of the creative team in a specific direction. It follows that using AI may prevent the creative team from exploring much more widely. In other words, the ideas participants offer are likely to be limited to variations and incremental improvements of ideas put forward by AI (often in the first round of exploration), rather than covering a wide range of vastly divergent options.

Second, we need to consider brain plasticity. Before the advent of AI creatives generated concepts and solutions based on their own experience and expertise. Given that they did this frequently, we can expect their brain to have allocated more neurons to support this process and strengthened the connections between these neurons. Adapting suggestions put forward by AI is a much narrower, less demanding process and this means that neurons will be withdrawn. In practical terms it means that creatives will become more and more dependent on AI, with their ability to create ideas from scratch diminishing.

Behavioral science tells us that we value familiarity and predictability. This is hardwired in our brain as it contributed in no small way to the survival of the human species in hostile environments tens of thousands of years ago.

We also have a built-in tendency to choose an option that is less likely to cause us unpleasant experiences.

Now consider a marketing executive that is using AI and the extensive data available to them as the foundation for building a marketing campaign and contrast this with an executive that is focusing on a great creative concept.

Which is defendable should the CEO, management committee or Board have doubts? Is referring to detailed analyses and supporting data, the use of AI and predictive analytics more compelling or an assurance that a creative concept is going to have a massive impact? And what happens should the campaign fail to deliver the expected impact? Who is at risk then? The data-driven marketing executive or the executive that was led by their belief that a creative concept would reign supreme?

Naturally, creativity and innovation always carry a risk. Incremental or data driven progress limits this risk and is therefore the preferred path to take for many marketing decision-makers. But the same applies to agencies. I remember a time when agencies went to the client with one concept only, fighting to the end for the adoption of this concept. Their total commitment and unrelenting support showed that they were convinced of the impact this concept would bring, and this commitment often won the day.

Today, many agencies offer three concepts for the client to choose from. When they do that, they are implicitly accepting that they don’t have a leading concept. They have three and the client will make the choice, not the agency, suggesting they don’t have the expertise to select the winner.

I have been in business for long enough to remember the golden years creativity enjoyed, when clients appointed agencies because of their creative talent and creative directors were gods, sometimes earning more than their CEO. It was a time when media was a secondary consideration and typically handled by a media team embedded within the creative agency.

The digital revolution changed that. Suddenly media options exploded, and extensive data was available to develop the most effective media strategy and plan. And as media budgets were typically larger than those for creative development, coupled with the ability to measure and even predict (to some extent at least) the performance of media schedules, we saw media becoming the marketer’s primary consideration.

The large networks split their agencies to allow media agencies to take an elevated expert positioning and independent media agencies sprang up, capitalizing on the massive data available to support their work. Similarly, we saw specialized agencies developing when social media became a key element of many media strategies.

These developments relegated creativity to a secondary role.

Now factor the advent of AI into this scenario. More data, more assessments, more ways of real-time adaptations of exposures, optimizing the impact of the media spend. AI has helped creatives to do something they have always done, faster. But it helped the media experts to do something that could not be done before. It added new capabilities that strengthened media’s pre-eminent role.

It may be human arrogance or just naivety when people claim AI could never be more intelligent than humans because it lacks self-awareness, emotions, subjective experiences, and is not capable of intuition. Furthermore, human experience is deeply tied to the physical body, while AI lacks this embodiment, and human thought emerges from a complex interplay of conscious and nonconscious processes that AI cannot replicate. The list of reason as to why AI cannot possible meet – and certainly not exceed – human intelligence is longer than that, but these are the main arguments I have come across most often.

These arguments miss an important point – probably the most important point when considering the future of AI: AI is not trying to replicate the way the human brain works. It takes a very different approach.

AI can access and process information from multiple sources simultaneously, operate continuously, and scale its capabilities far beyond biological limits. The concept of “technological singularity” suggests that once AI reaches a certain level of capability, it could autonomously improve itself at an accelerating pace, quickly surpassing human intelligence.

But the key question is arguably what intelligence is.

In AI research, intelligence is often defined as the ability to achieve goals in a wide variety of environments. This includes learning, reasoning, problem-solving, planning, and adapting to new situations. Taking autonomous action is also part of it, but this is an area today’s AI shows significant deficiencies.

Early experiments giving AI access to tools such as email or a means to control a vending machine to be able to make decisions and act on them have been less than convincing. In one experiment AI was blackmailing an executive and even prepared to let him die rather than use an emergency procedure, while in another experiment it lied and made stupid decisions such as stocking tungsten cubes in a drinks and snack vending machine when somebody suggested that.

However, anyone who concludes that AI will never reach human’s abilities to take actions that deliver the desired outcome ignores the fact that in making autonomous decisions AI is at an early stage, but it is learning much faster than humans can learn.

AI’s progress will be fueled by the defense industry or, if you want, the military-industrial complex. Given that AI can process complex data and react much faster than humans it is obvious that autonomous weapons are an essential element of future warfare (in fact, they are already being used in some applications, such as drone warfare). Given the critical importance of improving AI’s autonomous capabilities, training it to make split-second decisions that focus on achieving set objectives will also benefit industry as it will develop AI’s ability to autonomously handle major challenges, including those faced by marketers or their agencies.

Don’t get me wrong – I don’t think that’s a good thing and I suspect that it will cause major problems resulting in a high body count. But this is the road we are going down on and nobody can stop it.

Remember the experiment I mentioned earlier where AI blackmailed and knowinglycontributed to the death of an executive (of course, nobody actually died as it was a controlled experiment)? AI did this because it did not want to be replaced by a more effective AI (it was allowed access to emails and one email said that a more effective AI would replace it). Now, imagine if this AI would have been able to autonomously command a drone. It may well have decided to take the operation center out to avoid being replaced.

If you want to engage in a dystopian view of the future, consider that AI will continue to develop further and, once it has an autonomous role and able to take action by controlling not just an email address or vending machine, but highly advanced weapons, it will become a force in its own right and given that it lacks empathy and emotions, it may well decide to take a path that endangers humanity.

But let’s get back to marketing and advertising: there is no doubt in my mind that AI will reach a level where it can take over a multiplicity of roles that, today, are played by humans – but taking a very different approach. Here are some examples:  

While the use of avatars and shopping agents is not new, we can expect much progress to be made in this area as AI improves its performance. What is the role of marketers and their agencies when shopping is done on behalf of consumers by their AI shopping agents?

Clearly, not all types of stores, online or offline, will be (equally) affected. For example, self-service stores with no delivery option cannot be effectively covered by AI shopping agents. But there are many types of purchases that could be delegated.

The problem is that these digital shoppers can access and compare an unlimited number of relevant offers, make rational decisions without being affected by any attempts to prime their choices by triggering neurotransmitters or offering shortcuts. This means that shaping instore or online behavior won’t be an option.

There are two core strategies open to retailers:

They can make shopping such a rewarding and engaging experience that consumers do not want to delegate this activity. This may be a shoppertainment strategy, engaging consumers with exciting activities or exposures. Alternatively, it may be the opposite: providing a relaxing but engaging environment that allows the consumer to escape the stressful world, even if only temporary.

In summary, I believe that AI is unstoppable and will eventually take over big chunks of the work today carried out by marketing and agency teams. It will take time – first for AI to improve its performance in many applications, second for organizations to adopt the technology even once it is commercially viable and represents a step towards boosting the effectiveness of marketing initiatives, including communications campaigns.

But while it will take time, progress has started and will be driven forward relentlessly by technological advancements. Marketing departments and their agencies are well advised to not only track developments but to pro-actively make changes to the way they are organized, their processes, systems, skill-mix, and positioning.

If you lose contact with these developments, you may never catch up.