How AI is disrupting the world of advertising and marketing
Philippe Beaudoin is a venture scientist constantly looking for innovative ideas that can improve society. In 2020, he co-founded Waverly, a mobile discovery and curation platform that tackles the issues of disinformation, filter bubbles and addiction that are often associated with social media. Waverly uses a recommender engine based on natural language that lets users connect with their true aspirations, bringing transparency and control that are missing from other social media platforms. Before that, in 2016, Philippe co-founded Element AI, a world leader in operationalizing enterprise AI. He holds a Ph.D. from Université de Montréal, was a researcher at University of British Columbia and has spent the last 20 years honing his technological skills, including five years at Google. Philippe sits on various boards such as CEIMIA (Centre of Expertise in Montréal on Artificial Intelligence), the Montreal Science Centre Foundation and Montreal Digital Spring. He is a frequent participant in leading industry and academic conferences.
#1 - AI excels in its ability to identify the people most likely to respond well to an ad. What are the dangers of this data collection and processing?
There are many dangers. Data collection naturally raises privacy issues. What can be learned from the data gathered? What would be the impact if the data were disclosed to the public or shared with a third party who has different goals?
A danger that's discussed less often, however, and which I think is especially important — because it's insidious, pervasive and very harmful — is how people and society will be impacted by this new “ability to generate a reaction.”
As data science has evolved, it has led to an improvement in our ability to predict how people will react to certain content. Today, with artificial intelligence, those reactions can even be predicted on an individual basis. For example, we can predict whether someone will pay attention to a specific ad or whether they'll be inclined to share it.
These predictions give unprecedented power to the creators of advertising. They can actually keep developing iterations until they find the most compelling one that will grab the attention of the target audience and be able to influence them. That was already happening in the past — we all remember a catchy commercial from when we were kids — but now it can be deployed on a large scale.
Constant exposure to advertising that's almost scientifically designed to capture our attention definitely has a negative impact on us and our mental health. For example, we're hearing more and more about “doomscrolling,” where people spend hours scrolling through a content thread without being able to turn their attention away from it.
This ability to retain people's attention has given rise to what's now called the attention economy: an economic theory that views attention as a scarce commodity that can be deployed to influence people. Based on this theory, the richest people have the means to expand their influence more and more efficiently, which is a dangerous imbalance for society and even for democracy.
Without regulation, advances in AI will only continue to exacerbate these problems. It's up to all of us — and the advertising industry in particular — to raise awareness about these risks and encourage governments to respond to them.
Fortunately, there have also been advances in AI that aim to rebalance the attention economy by enabling people to be more intentional in how they consume content. The technology behind Waverly, for example, is a transparent recommender algorithm that lets people control their news feed to weed out content that manages to capture their attention without them wanting it to.
Research and technology solutions exploring those areas are rare and greater investment in them would be beneficial, both in terms of academic research as well as solutions developed by technology companies.
#2 - Will content personalization and automation make it possible some day to generate ads without human involvement?
I don't like to look too far ahead in my forecasts. There's nothing, a priori, that prevents an artificial system from achieving the capacities of a natural system. If we look far enough into the future, we can say “yes” to any question that asks whether a machine will one day be able to replace a human for a given task. In fact, AI in the future will depend much more on the legislation and policies applied at that time than on the capabilities of the technology itself.
But it's an interesting question if we look at the near future. Within a 10‑year timeframe, will AI be able to generate creative content without human intervention?
No, AI is unlikely to eliminate human involvement, but it seems clear that it will have a major impact on our creative processes.
That's what we can conclude from recent advances in content generation: GPT-3 produces very readable text and DALL·E 2 transforms short sentences into impressive images. In the coming years, we'll see more and more tools that enable humans to use those technologies in their creative processes.
The future therefore holds many surprises for artists, designers, animators, art directors and everyone involved in creative endeavours. They'll have access to new tools and software that will open the doors to new creative techniques. As has always been the case, evolving practices will lead to an evolution in the content created. No doubt that the human imagination will surprise us with how it uses those new capabilities.
We could also see a change in how creative work is organized. For example, the tools could supercharge the ideation phase by allowing a draft to be delivered in a few seconds. I predict that creative brainstorming will never be the same again.
Lastly, AI could make it much easier to create interactive content. Today, we associate interactive experiences with video games or, more rarely, certain expensive advertising stunts. However, advances in generative AI will likely reduce the cost of creating that kind of content. As a result, we could see the development of interactive ads that are able to dynamically adapt to a person or group.
Beyond the new creative horizons that will be opened up by advances in AI, we must be sure not to overlook the major impact they'll have on society. Those ads could become more compelling, more disturbing, more addictive, etc. It will be important to ensure that our legislation and policies take this into account to avoid misuse. The advent of the internet has buried us under a deluge of advertising, and I'm hopeful we'll be able to reverse that trend.
#3 - AI is also expected to become very good at reading our emotions or state of mind. How could marketers leverage that technology application?
It's true that AI is getting better and better at identifying an individual's emotions. There are various ways it can do that: - Analyze facial expressions captured with a camera. - Analyze the words written or spoken by an individual. - Analyze the behaviours of an application user: their clicks, scrolling speed, etc.
In fact, it's mainly because of these abilities to analyze emotions that the predictions mentioned above can be made. By detecting emotions, AI has a better ability to predict how a person will react to certain content. For example, if you spark someone's curiosity, you increase the likelihood they'll be attentive. If the emotion is anger instead, the person is likely to share or comment on the content.
In answer to the question of how marketers could leverage that application, my point of view is rooted in my values of fairness and respect for others. Don't exploit it! Or, better yet, use your expertise to speak out against the dangers of increasing our ability to detect and predict emotions.
Let's wish for a world where we can still be pleasantly surprised by creative content — advertising or other — without feeling that it's been scientifically created to grab our attention when we don't want it to.
Please know that cookies are required to operate and enhance our services as well as for advertising purposes. We value your privacy. If you are not comfortable with us using this information, please review your settings before continuing your visit.