As an emerging technology, AI faces and will continue to face its fair share of challenges. On the one hand, consumers remain wary about adopting new tech. Envisioning a world where humans are displaced by AI-empowered machines gone amuck may be haunting a few late adopters. On the other hand, companies express frustration that AI has yet to prove itself to be the magic pill that will streamline every business process and pave a path to bountiful profits. Here is tapping into the human side of AI.
In many ways, AI has been its own worst enemy.
AI has been operating under a mysterious set of rules only its authors, and IT understand, and the science has suffered a bit of an identity crisis. Like a gangly adolescent, AI is still trying to find its more significant purpose in the general marketplace.
The Medical Field
Everyone from medical researchers to web developers has been trying to find ways to utilize AI best. It may prove more difficult to bring AI into their fold, mold it for business purposes, and leverage its power. However, some believe that the process of AI hasn’t exactly been a fruitful one, so far.
Gartner predicts that through 2022, 85 percent of AI projects will deliver “erroneous” outcomes “due to bias in data, algorithms, or the teams responsible for managing them.” While that number appears extreme, it points to the real struggle enterprise is having to navigate the unchartered waters of AI integration into the broader business strategy.
Some business owners seem to feel it is very cool to say you’re in AI — but it’s another matter entirely to utilize the power of AI to serve business-specific needs.
AI and the Consumer
Consumers themselves aren’t faring much better. In a recent poll by a digital marketing firm, Blue Fountain Media, nearly half of the consumer respondents claimed they did not know what AI was and how it was being deployed around them.
These findings echo others found in similar polls that indicate there’s overall consumer mistrust of AI and a misunderstanding of how it’s currently being deployed.
Siri and Alexa
It is somewhat ironic that a consumer may mistrust AI — yet, if you ask them if they like their voice assistants, like Alexa or Google Home, many will say it’s been a great help in their daily lives. Are individuals not aware that voice assistance is enabled by AI? All smart home usage is also AI connected.
So where does all of this information leave us? Where are the next voice-assistant type successes that will find the elusive mix where both consumers and businesses see the benefit?
AI and the Human Element
The secret may lie in bringing a human element to AI. By working in a reverse-path from the wants and needs of the consumer back down through ways, an AI platform could service those needs.
The key to this new, mindful form of AI, is to focus on being aware and purposeful of the intentions and emotions we hope to evoke through any given artificial intelligent experience. The goal is to identify and articulate the core pain-points to solve and the positive value that mitigation of those pain points would drive.
An organization struggling to build meaningful consumer engagements, for instance, could identify the core issues that are driving the state of stasis and then decide how AI could be used as a means to unlocking it.
The mindful AI approach.
Using mindful AI to approach mental health issues, researchers at MIT were able to develop an effective AI platform for accurately diagnosing depression. Their neural network model analyzes raw text and audio data from natural conversations with a patient – which can detect words and intonations of speech that may indicate depression.
The AI platform is called context-free modeling, and it’s the first step to detecting mental illness just by using data received during casual interactions.
Applying mindful AI practices can also help identify and mitigate systemic biases entrenched in the raw data to ensure that AI system built on those data sources don’t magnify and further perpetuate those biases.
Garbage in/Garbage out.
Fundamentally, it comes down to the basic garbage in/garbage out precept, where the data sets for any given AI project are only as useful as the way they’re acquired and interpreted.
In their recent book Data Feminism, authors D’Ignazio and Klein make that salient point that data never “speak for themselves.” There are always humans and institutions interpreting and speaking for the data, which introduces biases and agendas that can affect AI outcomes.
Skipping the prejudice.
Getting “clean” data – that hasn’t been trained and tuned by prejudices – still remains a major challenge when it comes to constructing AI platforms that provide accurate results. At least at this stage, human activity continues to play a major role in the AI production process.
The entire AI process is why employing mindful AI techniques is crucial for achieving ethical and responsible AI models that augment human potential.
In a nutshell, applying a human-centric, mindful approach to AI comes down to this: Identify and focus on the needs of people first, so that AI applications serve human needs – and are aware of the human responses necessary to drive human progress.
Image Credit: Andrea Piacquadio; Pexels