©2019 Aigora

  • John Ennis

5 Tips for Sensory and Consumer Scientists When Vetting AI-Based Vendors


This article is part of Aigora's "Original Content" series, which consists of our original thoughts at the intersection of consumer science and artificial intelligence!


Disclaimer: What follows are my opinions and are not an endorsement or criticism of any service or company.

The difficulty with the term "AI" is that it sounds impressive but is, in fact, a broad concept that covers any use of machines to accomplish complex goals. Hence even technology as old as the Jacquard loom might be called "AI," depending on the marketing spin.  This ambiguity has created an environment in which false promises can thrive, leading to some in the computer science community to push back - see, for instance, Princeton’s Arvind Narayanan recent presentation “How to Recognize AI Snake Oil.”

In my day-to-day work, I'm often asked my opinion of this or that "AI-based" vendor. Hence, I thought it would be helpful to offer my opinions more widely - from someone deeply grounded in sensory and consumer science and who built multiple neural networks during his postdoc in computational neuroscience - to help decision-makers avoid poorly informed choices. As we have seen with Kraft Heinz, when management ignores the warnings of their scientists, disaster often ensues.

So, if you or your company are in the process of vetting an AI-based vendor to help with you sensory and consumer research, here is my advice:


Trust Your Gut



Evolution has equipped humans with excellent cheater detection systems. If, upon meeting a salesperson or founder with an AI-based offering, you feel uneasy or uncomfortable, listen to yourself. Your subconscious might be sending you a warning.


Remember That AI is Not Magic



There have been real and remarkable advances in the use of machines to accomplish complex goals in the past ten years, such as incredible advances in image and speech recognition, and it’s this fact that's fueling the current AI hype. To separate the reality from the hype, a good rule of thumb is that modern AI is excellent at: 

  1. Automating tasks that a human (perhaps an expert) can do quickly, such as recognizing whether a brand appears in a video.

  2. Tasks, such as searches through vast spaces, that a human could eventually do if they had enough time, motivation, and a perfect memory.  

While it's true there have also been advances in predictive analytics, especially when enormous quantities of data are available, it's also true that - as my friend Tom Carr says - you can't analyze information into your data. In other words, if the data are not high quality enough to contain the information you need, no amount of analysis will create the missing information. So, if there is a part of the technology that amounts to "this is where the AI magic happens," be skeptical.  

Remember That You're an Expert



Keep in mind that you, as a sensory and consumer scientist, are in the best position to evaluate the quality of sensory and consumer data. Speak up if you feel that the data involved simply don't contain the information required to fulfill the promises - it's your responsibility to protect your company from unwise ventures.


Watch out for the Razzle Dazzle



People who want to help others take the time to slow down and explain, in layman's terms, how their technology works. People who want to take advantage of others speak quickly, deliver memorized scripts, and make liberal use of both technical jargon and appeals to authority. If technical terms seem overly complicated, or if you just can't figure out how a piece of technology works despite repeated attempts, question whether you’re experiencing a “razzle-dazzle.”  


Ask for Help



For my final tip, I recommend that you leverage knowledgeable resources within your network to help you vet AI-based vendors, especially if you don't feel qualified to assess the more technical aspects of an offering. For this, you should reach out to your internal statistical, data-scientific, or IT teams, all of who would likely be happy to help. And, if you are at a large enough company that you have internal AI support, you should definitely involve them in your decision-making process.

Okay, that's it. Did I leave anything out? If so, please contact us and let me know! And, if you'd like to discuss any of these points in more detail, please feel free to schedule a discovery call.

If you'd like to receive email updates from Aigora, including weekly video recaps of our blog activity, click on the button below to join our email list.