Why advertising urgently needs its own code of AI ethics
The world is abuzz with the potential of artificial intelligence to transform every industry: from smart cities to healthcare and autonomous vehicles to reaching the final frontiers of space. AI is going to impact every single industry and everything we do as a society. Next to these stories, perhaps pondering the future of advertising feels almost trivial.
But it is in our technology-enabled and AI-enhanced ways with which we are now able to influence large (or small) swatches of populations, communities, interest groups and individuals that should give us brave reasons to pause and ponder. Marketing with intelligence gives us the potential to produce outcomes that are more relevant and profitable.
AI is allowing us to connect the most motivating messages with the most motivated individuals. Last year alone, Essence was using machine learning to beat expert sales forecasts 70% of the time; match or outperform our best human optimizers nine out of 10 times, and even eliminate ads that would have otherwise annoyed consumers and wasted money.
"As one of the first industries to wield this technology, we need to wrestle with its implicit ethical dilemmas, implications, and responsibilities now, while we have time to align on the codes and standards for how it should be used."
This ability to communicate at scale with relevance, personalization, and potency is giving marketers enormous new powers. Yet, as one of the first industries to wield this technology, we need to wrestle with its implicit ethical dilemmas, implications, and responsibilities now, while it is still in its infancy and we have time to align on the codes and standards for how it should be used.
The first question is: how do you code morality into algorithms? Advertisers are required to make moral judgments about whether their ads meet subjective standards for being “good”, “honest” and “decent” on a daily basis. This may not be as dramatic as coding for who an autonomous vehicle should prioritize in the advent of an accident – child or an elderly couple, for instance – but it is as nuanced as it is complicated.
"The first question is: how do you code morality into algorithms?"
Whose duty is it to protect the gullible from conspiracy theorists; the disenfranchised from polarizing news cycles; recovering alcoholics from an effective new alcohol campaign; or chronic gamblers from a gambling den’s welcome message? The more you think about these scenarios, the more you can go into a dark, depressing and scary rabbit hole. Who and what will be the vanguard of humanity’s moral compass, and how will it be governed? Algorithms could conceivably pair even the most harmless platforms, brands, and content into terrible and exploitative combinations to users at a scale where complete oversight is virtually impossible, and we are seeing this already.
We need a governance protocol that is infinitely better than today’s brand safety mechanisms if we are enabling machines to make these decisions for us. Beyond codifying morality lies the difficulties algorithms have with incomplete information. Humans are able to consider context and weigh outcomes to make decisions when we are uncertain. Algorithms resolve uncertainty through experimentation, making thousands of trials and errors to find an approach that works.
"We need a governance protocol that is infinitely better than today’s brand safety mechanisms if we are enabling machines to make these decisions for us."
To get to the right decision, advertisers will either need to accommodate an unprecedented amount of error, which is in itself problematic, or give algorithms access to an unprecedented amount of contextualizing information, which can be easily prone to manipulation and biases.
The questions AI raises for advertising do not end there. It is obvious this new era demands new levels of accountability and leadership – but from whom? This responsibility, of course, belongs to all of advertising!
In recent months, the leading researchers of AI have begun to converge around principles that will help technology companies avoid creating intelligent systems that reinforce the inequities and biases of today. But to date, advertising has largely been excluded from this conversation.
Companies such as Google, Facebook, Amazon, and Microsoft have each taken important steps to publish important guidelines for the development of AI, and now the advertising industry must follow suit. As long as technology companies remain reliant on advertising to fuel their business growth, it is the duty of advertisers to have a voice in shaping the code of ethics that should govern us, universally.
This can sound utopian, but so is the very topic of morality. What we know is abundantly clear – the intelligent era is upon us, and the industry as a whole needs to come together to form a code of ethics to govern the use of AI. We all have a big role to play, to boldly go where we are all headed, together.
This article, co-authored with Kunal Guha, has been adapted and was originally published by Marketing Magazine.