Why we need to move away from anthropomorphic naming conventions in AI

AI has emerged as a focal point of discourse and, at times, vigorous debate. The underlying concerns are not unfounded. Its infusion into nearly every sphere of work and lifestyle has touched the core of human existence. The technology, while propelling a surge of innovation, optimization and a host of other beneficial applications, has also broadened the horizons for troublesome implications with its widespread adoption. 

The spread of fake or false information on a large scale has significantly increased. Biased and discriminatory programming is another concern, as are issues related to consumer privacy, identity fraud, the rise of AI capitalism, job displacement, economic inequality, AI hallucinations, unethical values in AI systems and the use for illegal purposes, among others. The list goes on. 

Each of these implications has not been lost on experts. They have discussed such scenarios in public forums. While efforts are being made to minimize or altogether eliminate risk, experts admit that they are ultimately engaged in a guessing game. Anticipating how AI might evolve could very well be outpaced by its actual evolution. The very public debate between the two godfathers of AI, Yann LeCun and Yoshua Bengio, has not helped matters, leading to further confusion and anxiety. 

However, an area of concern that has garnered little to no attention is one that has been top of mind for those in the innovation and marketing industries: When developing products and services, how should one go about naming or branding something new? 

What’s in a name?

While the naming of technologies might seem trivial, it could have unforeseen consequences in today’s world, especially concerning the anthropomorphic conventions often used in naming AI agents, bots and the like. 

Take service interfaces, for example, particularly those in the Western world, which often lean toward female-gendered names. McKinsey’s AI assistant is called Lilli; Hanson Robotics’ social robot responds to the name Sophia; Microsoft’s personal activity assistant is known as Cortana; and then there’s the most famous virtual assistant of them all: Alexa — although “she” may be tied with Siri when you’re on the go. 

All that to say, female-gendered names given to AI bots that you control can lead to confirmation bias and reinforce the idea that women are subservient to men. This is not merely an implication. Furthermore, the use of anthropomorphic conventions in naming AI can affect its perceived potential to harm, as it begins to appear to have a “mind of its own.” 

Moreover, behaving like a God without fully understanding the potential consequences of what you’re creating carries the risk of total resentment towards the technology, should its intelligence surpass that of humans. While this may seem trivial, it could facilitate the fear-inducing transformational approaches that some tech companies have employed. 

Where do humans fall on the spectrum of value if this technology can surpass anything someone can create? For many people, human-like AI makes it difficult to differentiate between the non-commercial and commercial pillars of human-to-human relationships, such as personality, morality and trust. 

Experience: The name we give ourselves

With the first wave of digital transformation, there were key learnings about its second-order effects. As an example, one only needs to look at the mass commercialization of social media and digital interfaces. Due to the loosening of social norms, society has begun to see an acceleration of anonymized aggression and even hate speech online. It’s still too early to determine the long-term implications, as digital anthropology is a relatively new scientific field. However, repercussions are likely and could intensify over time. 

As inconsequential as it may sound, the same “rule” applies to naming AI in all its various iterations. The technology should be viewed as an enabler rather than competition for real individuals. Innovation and marketing teams must help their clients better navigate the AI sphere, and part of the journey will involve the naming conventions employed. These conventions may very well determine the long-term success of new technology. 

As Martin Heidegger once said, “Everywhere we remain unfree and chained to technology, whether we passionately affirm or deny it. But we are delivered over to it in the worst possible way when we regard it as something neutral; for this conception of it, to which today we particularly like to do homage, makes us utterly blind to the essence of technology.” These words are something to always keep in mind. 

A rose by any other name

For guidance and confirmation, you can look at the naming standards used for pharmaceuticals and those used for URLs. 

Naming regulations are in place to ensure that no one can make assumptions about efficacy, which can heavily affect consumer perception and the likelihood of a purchase. Regarding naming generic drugs, they must use two syllables in the prefix, be absent of specific letters and avoid medical terminology. Names that are too fanciful may also invite greater scrutiny and lead to rejection. 

With URLs, the emphasis is on keeping things clear and concise. Another recommendation is the use of lowercase letters and avoiding special characters. The goal is to have a name comprised solely of letters, hyphens and numbers, along with “names” that specifically explain the site’s components. Deviating from these guidelines can cause problems. 

Similarly, AI tools would benefit from similar naming conventions. This approach helps keep the technology, at least in the eyes of the consumer, squarely in the product category. After all, that’s what it is. Anthropomorphic naming conventions can do the opposite, making AI appear to be a substitute for humans. Moving away from giving the technology a human name also opens the opportunity to explain the objectives and capabilities of that technology. 

Names are now brands. Truth be told, they’ve been brands before branding was even a thing. Richard Branson serves as a testament to this. Celebrities and public figures like Cher, Madonna, Beyoncé, Taylor Swift, Elon Musk, Marie Kondo, Rihanna, Neil Patel and Simon Sinek have also become brands in their own right.

Not that people would necessarily confuse AI for a real person, but using anthropomorphic naming conventions can lead to implications far beyond gender stereotyping. It can cause people to treat technology as human beings, beginning to blur the lines between human and machine. 

Katrin Zimmermann is CEO and managing director at TLGG, an Omnicom Precision Marketing Group (OPMG) company.