“Alexa, remind me to take out the trash.”
“Hey Siri, play some music.”
“Cortana, tell me what the weather’s like.”
Man! I hope your friends respond politely to your slew of demands.
Nowadays, these asks are unmistakably requests to a voice-enabled home assistant, capable of streamlining your routine with conversational AI capabilities. Over time, AI tools have evolved to fit almost seamlessly into our daily schedules as humanoid helpers, mirroring the way we talk and interact.
As your organization embarks on new data science and AI initiatives, you may find yourself drawing comparisons between the intelligence you are developing and that of a personified AI device you have encountered or hoping that one day your product will have a cool first name to show for itself, too.
Modern AI Applications & Devices
Take a second to consider all the products you have come across with intelligent capabilities that attempt to approximate human-to-human interaction- your friendly voice-controlled speaker, your latest DMV online chat experience, or even a chipper paper clip making suggestions as you type up a document.
Consider the benefits you have gleaned from these tools and how these benefits stem or stray from the AI’s human-like qualities. Why does an AI agent have to be humanoid to be helpful? Is more human always better, and how can this idea of AI personification impact perception and success of AI development in your business?
Microsoft Paperclip Office Assistant – Image via Geekwire, 2019
Understanding Human Instinct
It is straightforward to understand why human qualities are intentionally strived for in intelligent technologies- in short; humans love humans! When we refer to a machine in the same way we refer to our friends and families (with a real name) and interact with it similarly, we frame it in an affectionate light and are naturally more likely to have a good first impression of the tool.
People innately react more warmly to personified objects over lifeless machines or systems. This phenomenon is potentially fostered further through expectations of AI we might carry from past experiences- say, maybe watching a few too many sci-fi movies. It feels fitting and satisfying that AI has a name, voice, or even personality. The initial impression left by AI might be more positive if the tool engages using human characteristics, but how can these qualities impact the overall reception and adoption of AI later down the line?
Let us consider expectations around AI to explore why the personification of an AI tool might be more harmful than helpful. When AI is presented as human, we expect human intelligence– well rounded, flexible, holistic intelligence- and this mindset can put a damper our experience with AI before it even begins.
Framing AI in this way can predispose users and give the impression that the AI is falling short, regardless of intended performance, or can even motivate users to push the limits of the tool (i.e., attempting to order a pizza from an automated technical support line). Users have a better chance of successfully troubleshooting any issues or adapting the way they interact with AI when it’s clear that conflict resolution won’t fare. Adding “please” when asking Alexa to turn on the lights I never connected, unfortunately, does not seem to accomplish much!
In short, AI is likely best received when expectations are measured and based on true capabilities, not a futuristic fantasy.
Image via TechTarget, 2020
Misrepresentation of AI in an Organization
AI personification can also negatively influence perception of the tool not by falling short but by overstating or misrepresenting the goal of AI in your organization. The idea of an AI agent perfectly mimicking a person, being equipped with intelligence akin to a human’s, and carrying out complex work with no intervention isn’t just a little creepy; it’s unrealistic and can paint a false picture of AI intending to put people out of work, take over our lives, embark on world domination, etc.
Outside of the movies, AI is a front of innovation capable of improving the way we do our jobs, providing real solutions to our problems, and enabling us to reap efficiencies impossible before. AI does not need to be human to fulfill human needs when the true goal is to empower, support, and serve people- not replace them.
Finding Value in AI
All in all, don’t sweat it if your AI tool can’t sing and dance or pass the Turing test– the heart of AI is simply the use of data to make human lives easier and better, not robotic speech patterns or animatronic blinking. Take care to frame and market your data science and AI initiatives in a way that is honest and realistic, and let your findings and results speak for themselves. Nothing is truly more exciting, innovative, or transformational than beginning to put your data to work for you.
Whether you’re developing a complex machine learning solution or considering where to begin on the data science and AI front, we’re always here to help. For additional information on our solution to build actionable machine learning insights, get started with 3Cloud today.