How Do We Make Marketing Replicants a Reality in 2049?

Share this post

Blade Runner returned to the big screen last month and with it came a new generation of artificial humans – or replicants – more obedient and long-lasting than their predecessors, but still hard to distinguish from their creators. Set three decades from now, Blade Runner 2049 is both fictional and unashamedly futuristic. But will technology realistically advance to the point where replicants can be created in 30 years’ time?

The marketing industry relies heavily on technologies such as artificial intelligence (AI) and it is these developments that provide the key to creating replicants. While marketers don’t necessarily need artificial humans, they do require something similar – a highly intelligent system that can analyze extreme quantities of data and react instantly to changing environmental parameters, optimizing campaigns on a continual basis.

Recent advances in AI, such as machine learning and deep neural networks, allow machines to replicate small elements of human-like behavior including recognizing pictures and patterns, or even – in the case of IBM’s Watson – composing music. If a human process can be codified, it can be replicated in AI and performed much faster than by humans across larger data sets.

But so far AI can only perform the specific processes it is manually trained to undertake. For instance, DeepMind’s AlphaGo can beat the world’s best Go players, but it can’t currently apply its knowledge to other board games – or indeed optimize a marketing campaign. To be a realistic possibility, Blade Runner style replicants would require artificial general intelligence (AGI), allowing them to perform all the intellectual tasks of a human brain – this is a very different proposition to the AI used today, and requires a great deal more research and development.

The marketing industry finds itself in a similar situation. While significant advances have been made to incorporate machine learning algorithms into advertising technologies – with those ad tech companies actively using AI enjoying significant revenue uplift – the industry is still only at the start of the road to AI. What the marketing industry is doing now is no more than applying advanced statistics at scale – for example to leverage data insight to deliver individualized marketing across campaigns. To realize the full potential of AI – making both replicants and their marketing equivalents a realistic possibility – there needs to be rapid evolution in three key areas:

Boosting data volumes  

Nobody knows exactly how much data the human brain holds, with estimates widely ranging from 2.5 petabytes (2.5 million gigabytes) up to an incredible 500 petabytes. But suffice to say, it’s a massive amount – trained by many years of life experience and far more of evolution. To make the leap from the AI available today to a level where the human brain can be fully simulated would require a colossal increase in the volume of training data used for development – and the picture is similar for the extreme data needed to optimize advertising technology.

Nobody knows exactly how much data the human brain holds, with estimates widely ranging from 2.5 petabytes (2.5 million gigabytes) up to an incredible 500 petabytes.

A lack of capacity to handle extreme data means the models used in marketing are based on just a few days’ information. But this is insufficient for statistical accuracy. Factors such as unusual weather conditions or political events can easily sway data for a particular day, week, or month, meaning it won’t be representative over a longer time period. To achieve a holistic, seasonal, global view and far greater accuracy, advertising models should be based on several years’ worth of data rather than just a few days.

Access a variety of sources

It’s not just the volume of data that must increase moving forward, but also the variety of data types and sources. The data inputs required to simulate each of the brain’s hundreds of billions of neurons are not yet fully understood, but fortunately for marketers there is already a much clearer view of the data needed to drive intelligent marketing.

Rather than just relying on response data, marketers must include other information types such as temporal and geospatial data to gain a fuller picture of the consumer journey. They can use a variety of data streams from smart cities, beacons and sensors, mobile devices, and the Internet of Things, and can draw insight from this data to truly understand the consumer’s in-the-moment needs and respond accordingly.

Accelerate processing speeds

The future development of AI depends on what can be achieved in processing capacity.

The future development of AI depends on what can be achieved in processing capacity. Development of the enhanced use of AI for marketing requires machines to absorb massive amounts of data, from a wide variety of sources, understand the immediate situation, and respond instantly with the most appropriate action.

The human brain is naturally able to operate like this as it employs massively parallel distributed processing, meaning it can run multiple processes concurrently rather than sequentially. The ad tech industry is already making advances in processing – for instance by using GPU-based architecture – but there is still a long way to go to gain the processing power needed to realize the full potential of AI.

In reality, marketers are unlikely to find themselves working alongside human-like replicants by 2049. But as long as they can increase the volume and variety of data they are using as a basis for campaigns, and the necessary developments in processing power and analysis tools occur, marketers are highly likely to enjoy the equivalent – highly intelligent AI-based marketing systems that analyze extreme data sets in real time, allowing marketers to instantaneously take the most effective action to deliver more impactful ad campaigns.

Share this post
No Comments Yet

Leave a Reply

Your email address will not be published.