Considerations for Technologies That Are Designed for Everyone
Artificial Intelligence is changing the way we do…well, pretty much everything. It’s been slowly creeping into our lives over time, and is now a major player in our daily routines.
As a society, we’re quick to embrace new technologies, and while on its face technology seems relatively harmless in its quest to solve our problems and make tasks easier, at the end of the day it is designed for humans–and all humans are different.
Exclusive and Inclusive Technology
There has been more attention being paid lately to the idea of exclusivity in technology, an issue on which Microsoft took the lead last year with its call for public regulation and corporate responsibility for facial recognition technology.
While the intentions and capabilities of AI can be exceedingly beneficial, it can pose significant challenges in regard to privacy, public policy, safety and basic human rights, including certain biases of race, gender or socioeconomic status.
That issue is far broader than we will cover in this post, and as Medium notes in a recent article, addressing it would require more diverse datasets that represent every customer, and any bias in the system makes it more difficult to collect that data as those underrepresented individuals have little incentive to participate.
However, on a higher level, it is the need (and implicit desire) to make AI work for everyone that has made us consider how the more traditional rules of branding can apply to today’s burgeoning technologies. AI technologies are targeting widespread audiences, and making it inclusive and effective means understanding who it’s for and how they’re going to use it.
“Recommended for You”: Making Sure Behavior-Based Recommendations Make Sense
Love it or hate it, one of the key functions of AI is to predict what we want based on our browsing history, purchase decisions, and even conversations. Different generations tend to have varying opinions on these capabilities (which goes back to issues of privacy and human rights) but at a basic level, the ability of AI to tap into our preferences is similar to content marketing and marketing automation tactics. As digital marketers, we do a lot of work with predictive analytics to feed content and information to users at the right time and in the right way based on their behavior. In this way, we ensure their website experience or interactions with an email take them further along their “buyer journey” with useful information at every turn.
Implementing predictive analytics into marketing activities relies on user demographics and behavioral information–just like AI–and runs the same risk of losing a user’s interest if executed incorrectly. We all know how irritating it can be when Google delivers a recommendation for a product we either already purchased or have no interest in, or, conversely, how great it can be when the video game we looked up recently pops up with a discount code. It’s imperative for AI to understand its audience’s behaviors on a deeper level to make that gap between humans and technology almost imperceptible.
Never Underestimate the Importance of Research
No brand was ever successful launching a product without any knowledge of the audience it was designed to serve, which is because consumers have expectations for the brands with which they engage. Yet, there seems to be a perception that AI is just built smart and will always deliver on our expectations because “it just knows.” This scenario, however, is far from accurate. Design teams, engineers and the public need to focus on research as the foundation for effective technology experiences, and no one is better suited to train AI technologies than humans themselves.
It’s clear to us at this point that the intricacies of human behavior are nearly impossible to emulate exactly with technology. There are nuances to individual human beings that go far beyond the capabilities of artificial intelligence. However, with the right research, input and training from target users, AI can be smarter, more intuitive and more successful.
Brands Turn to the Crowds–AI Can Too!
Much like the need to conduct research, brands also can benefit from leveraging actual customers to use their product, provide feedback and ultimately help improve the design based on real-life use.
Crowdsourcing works well as a tool for brands, and it has also worked well for AI. Take Mozilla for example. Its initiative, Common Voice, is crowdsourcing voice data for its machines so voice recognition can be more accessible to everyone. This approach allows real people to be part of something driven by AI, thus injecting the human element into the technology in a more effective way. Rather than a team of designers and engineers sitting behind closed doors with no reality-based understanding of the implications of their technology, information is fed into the technology directly from people representative of those who would actually be using it.
The continued evolution of AI technologies points to an exciting future for humankind, but also one that could be fraught with challenges around how these technologies impact society and individual users. From a brand marketing perspective, the path to success for AI should be rooted in a deep understanding of all target audiences and the markets for which it is being designed.
AI technologies are products like any other, and when they resonate with the personalities, lifestyles, demographics and behaviors of the target audience, they will be all the more successful in delivering that human-centered experience we expect.