What’s not to like about Maltesers - the all-time favourite snack that is associated with good times at the cinemas? But the last year, with the shutdown of picture houses, more people in the UK took to baking. Working with Google, Maltesers deployed machine learning that created an AI-optimised cake and cookie recipes using chopped and whole Maltesers complete with an all-original Malteser buttercream topping.
Lip-smacking cake recipes aside, AI is already being used to identify genetic mutations within tumours, will someday solve future pandemics, predict tsunamis and other natural disasters and potentially learn to fight cancer by detecting cancer cells. All this is made possible by scientists who are implementing deep learning techniques to build a simulated neural network that can compute algorithms and predict results like a human.
AI in digital marketing
Vast amounts of data accumulated over the past decade or more have enabled the advertising industry to launch, scale and measure digital campaigns by correlating myriad data sets for predictive analysis. COVID-19 has further accelerated digital adoption and automated a number of human processes.
IPG’s Reprise, for instance, has been using AI to write engaging pieces of text that are indistinguishable from human copywriting. The same deep learning technology is also producing high-resolution, photo-realistic pictures of fictional humans or objects that do not exist in the real world. “AI lets our creative people be creative and takes away the burden of repetitive work while serving as an inspirational sidekick during brainstorm sessions,” says Vincent Spruyt, global AI officer at Reprise.
AI has also seen the growth of digital influencers such as Lil Miquela and League of Legends’ Seraphine who are being used for their ability to connect with younger audiences and being omnipresent across geographies. In November last year, Lil Miquela signed a US$ 10m deal with Creative Artists Agency. That is not all - there is a “fuccboi” Blawko, AI supermodel, Shudu as well as AI-driven virtual girlfriend Xiaoice who is romancing digitally connected yet socially isolated young men in China. The bot has already reached 600 users, which is nearly half the population in China.
In the Asia Pacific region, Mindshare Australia’s cricket AI called Monty - the world’s first "predictive commentator" - is cited as one of the best and foremost examples of AI-led digital campaigns. Launched in 2018, Monty could tell when a wicket was about to fall and telecast it via pre-roll, digital out-of-home, Fox Sport app and real time ads. Monty was even on Google’s Assistant.
The campaign for Foxtel combined data from Google and from UK-based Opta to analyse 83 variables including stadium location, weather, moisture, pitch condition and team dynamics for every ball over a year’s worth of game by applying AutoML alphas. “We named the AI commentator Monty, as the first wicket it accurately predicted was of England’s left-arm spinner, Monty Panesar,” said Brooke Thompson, strategy director at Mindshare Australia, who explained that it was done with only one objective in mind: to enhance the consumer experience.
Enhancing consumer experience
Those in the ad industry know only too well that the best marketing campaigns place consumer experience at the heart and centre; AI’s objective should not be any different.
The only difference is, for AI campaigns to create similar experiences requires access to and training on data sets at an industrial scale. Companies that have already transitioned to digital such as those in banking, finance or travel industry will be more advantaged. But that doesn’t mean others are not. The three technology giants, namely Google, Facebook and Amazon alone can provide off-the-shelf data stack for companies to leverage.
However, as technology companies respond to a privacy-first world, mapping, targeting, connecting and measuring various data sets will mean a focus on developing a robust first-party data.
Michelle McGrath, CEO of LENS10, a MightyHive company, said : “Marketers should set up their foundation by collecting clean, reliable data with appropriate permissions from multiple connected sources; test and learn by playing with the cutting-edge tools and techniques to identify opportunities to generate outstanding success.” McGrath added that the success metrics of AI-based marketing campaigns should be measured against scalability, speed and longevity to ensure the campaign is future proof.
With the general adoption of AI, the digital industry can expect automation of low-risk processes; ad-ops being most ideal for implementation. According to Graham Wilkinson, product strategy and innovation executive at Kinesso, the marketing intelligence engine of IPG, most companies will eventually use a level of machine learning in the future because it will come pre-built into our systems. "At the lower end of the equation, AI has a potential to become commoditised, which shouldn’t be too worrying," Wilkinson said.
He however cautioned that with the rise of General Data Protection Regulation and California Consumer Privacy Act, AI making unethical decisions would be detrimental and the ad industry must be careful about the potential risk emerging from data privacy or unfairly putting one group at advantage over others.
Ethics of AI in the ad industry
“When machines do a job too well, they can take away livelihoods and put humans at risk,” Wilkinson said. “We built this industry and have a duty to take care of the people whose livelihoods depend on it,” he added.
Alluding to the Netflix documentary, The Social Dilemma, Wilkinson warned that he would not want to see those who built the industry feel remorseful of the monster they have created. “We must continually ask: will AI place humans at risk? Will it put one group of people at an advantage?” says Wilkinson who works to build parameters that will eliminate negative biases as well as parameters that acknowledge the positive biases and moderate the two.
Mindshare Australia too feels strongly about the ethics of AI. “The onus sits with us,” Thompson said. “We have a responsibility to enhance the consumer experience, not exploit it. Don’t jump on the AI bandwagon just for the sake of it,” she warned.
The agency, which built Monty in 2018, has not launched another campaign since. “Monty was conceptualised and launched purely by keeping consumer experience in mind. As such, there was no hurry to create another one for the fun of it,” Thompson explained. To guide their teams, Mindshare has even developed a detailed guide on the ethics in AI.
IPG too has a fair and equitable guide for its agencies. Given the power of AI today, Spruyt believes that most, if not all, applications of AI should take a human-in-the-loop approach. “We should do everything we can from a technical perspective to yield fair and ethical models,” he said. “For example, an AI-generated piece of content should always be reviewed by a copywriter and should never be pushed to a live campaign automatically, not because it wouldn’t work, but simply because we have the responsibility of being accountable, transparent and fair," he added.
AI serving a human purpose
Google CEO, Sundar Pichai has stated that AI was more profound than fire. Tesla founder Elon Musk proclaimed in a 2017 CNBC interview that AI was more dangerous that North Korea. The two leaders’ comments show that AI offers possibilities but its power can’t be underestimated.
It is apparent that AI can’t be left to its own means. But what is also apparent is that the ad industry’s AI narrative is moving beyond just finding solutions to managing risks - whether it is removing unconscious bias, maintaining job security of its staff, preserving user privacy and ensuring ethical decision making.
With the big technology platforms making AI accessible to data scientists, the ad industry knows that to build trust in AI, it can’t default to government regulations. Instead, it must make fair and equitable use of technology that balances acceleration with ethics. For the ad industry, human purpose trumps all, and it is steadfastly committed to it.
The writer is Madhavi Tumkur, PR director and head of marketing at Enterprise PR and Marketing.
Photo courtesy: 123RF