Without a strong understanding of machine learning, it’s easy to be swept away by headlines that read “Facebook engineers panic. Pull plug on AI after bots develop their own language” or “You Will Lose Your Job to a Robot — and Sooner Than You Think.”
With a newsfeed flooded with articles predicting a dark future for anyone without a metallic heart or algorithmic brain, it’s hard to look the other way.
To challenge this flawed depiction of our future, Katherine Bailey dedicated her Acquia Engage session to weeding out some misconceptions around artificial intelligence to get to the bottom of how machine learning tactics can make a positive and tangible impact.
AI has become a blanket term often used to describe a range of emerging technologies. In the process, our understanding of what AI is, and can realistically do, has blurred. As the principal data scientist at Acquia, Bailey sought to bring the real potential of machine learning and more reasonable concerns to the forefront.
The Issues that Matter
Machine learning is a specific set of techniques that enable machines to learn from data, and make predictions. A primary concern for many people is the idea of singularity: the point in time when machines reach a higher level of intelligence than humans and, in turn, take over the world (essentially).
In reality, this is nonsense. There are plenty of things to worry about with the future of machine learning, and the singularity is not one of them.
What’s more of a cause for concern is the machine learning bias we see on a regular basis. For example, there are certain systems that claim to predict your interests based on your name, and some that conclude “man” is to “computer programmer” as “woman” is to “homemaker.” Even worse, some claim to predict terrorists based on facial features. Clearly, there are major flaws with the methodology powering these predictions, and these issues should be addressed.
Machines only learn from the data at their disposal. When the biases of our past and present fuel the predictions of the future, it’s a tall order to expect AI to operate independently of human flaws. If we want machine learning to affect our lives in productive and ethical ways, awareness of this fact is critical.
The User Experience Challenge
Without even realizing it, people interact with machine learning systems on a daily basis, surfacing a whole new set of issues for consumer-facing brands. Google’s “did you mean?” feature was built to help users in the search process, but in practice, there’s no denying it can get a little creepy and feel invasive. On the other hand, if you’re given a Netflix or Spotify recommendation that’s off base, users get frustrated and expect better understanding of their preferences. Today’s companies are forced to toe the line between helpful and creepy.
Companies are also running into problems around AI’s inability to explain its suggestions and recommendations to users. For example, Alexa can’t explain her reasoning for recommending a new product or service. Although we want to trust the devices that give us answers, at the same time, we want to understand where they come from. Solving these user experience challenges is what will separate the leaders from the laggards in the machine learning realm.
Baffling with Brilliance
With big picture issues and UX challenges not letting up, where do we go from here? We need systems that understand context and how the world works, and this isn’t going to be solved overnight. For now, tech leaders need to look past the hype and the bafflement, recognize the roadblocks right in front of them and set more achievable goals.
Clearly, there are obvious limitations involved with AI and machine learning. Mastering how to work within these bounds, building safeguards against machine learning biases and doubling down the UX challenges that affect our daily lives are keys to building a brighter future.
To watch Bailey and other Acquians discuss how machine learning fits with customer journey's click here.
To view Bailey’s full Engage session, watch below: