This Startup Wants to Build Artificial Intelligence With a Heart was originally published on Springboard.
When Spike Jonze’s film “Her” hit movie theaters in 2013, an awkward buzz around the unknown novelty of artificial intelligence was just beginning to snowball into full-fledged paranoia. Critics went so far as to title the film the scariest movie of the year, casting AI as the ultimate despot that left us—the viewers and subsequently unwitting characters of our own dystopia—“enslaved not to robots but corporations.”
Since the shift toward that mentality nearly a decade ago, dispelling confusion and fear around AI and machine learning has proven to be a challenge, one that Maria Dyshel and Hobson Lane, the co-founders of the startup Tangible AI, are dedicated to overcoming.
Tangible AI offers AI consulting services and helps organizations use AI in ways that result in a positive social impact. Both Dyshel and Lane cite “Her” as a much-too-common reference point for the public understanding of artificial intelligence. In the early stages of establishing the startup, both Dyshel and Lane found themselves in conversations that time and time again brought up the Spike Jonze film, and along with it, a fear of AI as a binary thing that would “either save or destroy the world,” said Lane, who is also a Springboard mentor.
For Lane, artificial intelligence is far from frightening. In fact, Tangible AI’s aim is to make sure that it becomes the opposite.
Making AI work for us
For Dyshel and Lane, the moment people can get past their fear of AI as a creepy, convoluted algorithm or a scary seductive robot whose female voice reinforces gender bias, it becomes something much simpler: a tool that can be used for good.
Tangible AI has identified three primary pillars as a way to make AI accessible and affordable: AI consulting, chatbots, and data science and machine learning. Dyshel and Lane are now working to show social organizations ways in which AI can be better used to help accomplish tasks like engaging people around a cause or automating processes in industries like healthcare and education. The startup’s clients are mostly nonprofits and organizations that support disadvantaged groups. “AI is not science fiction. AI is here, and everyone should be using it—particularly people who are doing the important work in the social sector,” Dyshel said.
Take education, for example. Earlier this year, Tangible AI partnered with FineTune Learning, an ed-tech startup building online education platforms, to build an app that helps students study remotely, with assignments easily accessible on multiple devices. Another app that Tangible AI is working on is powered by an AI assistant that allows students to ask specific questions (as they would a search engine) without getting back answers full of misinformation or bias. (If for example, you were to Google “Where was Barack Obama born?” the results would likely include answers from news organizations, sponsored articles, and the like; a Googler would be hard-pressed to end their reading at “Honolulu, HI.”)
The long-term consequences of the spread of misinformation inspired Dyshel and Lane to create a safe place for students where they can learn by querying for factual information, and a bot can reply succinctly—without any extraneous information.
AI in the time of Covid
Just as this year’s Covid-19 pandemic expedited worldwide demand for tools to enable remote learning and working, so too did the need for creating the feeling of interaction and connection when it’s physically impossible. This trend to acknowledge our more basic instincts of socializing has created a reemergence of a corner of technology that almost feels like a trip back in time: the chatbot.
Tangible AI is currently beta-testing a new, free-to-use chatbot named Syndee, which was designed to help women recognize and overcome feelings of imposter syndrome. Syndee begins by asking about the user’s familiarity with the concept of imposter syndrome, provides some information about what it is, and then gives the user space to share specific examples from their own experience. Syndee then asks more tailored questions designed to help the user identify unhelpful and unhealthy patterns of thinking and behavior, before finally offering a series of actionable tactics and exercises designed to help break these patterns.
While they might seem regressive, chatbots are an excellent step towards democratizing AI. While there are millions of households without smartphones—a 2019 study showed that 19% of Americans do not own one—the majority of those households are concentrated below the $30,000 income threshold, in which the percentage of adults without smartphones jumps up to 29%. “We’re going retro to help those people who don’t have access to smartphones to get an education—with the hope of an eventual global deployment,” Lane explained.
For Dyshel, the need for high-functioning chatbots became apparent when working with social organizations in less-developed countries that offered essential in-person services—and the pandemic making those services near impossible to provide. “Suddenly, these organizations started to realize that they need to provide those services digitally—and we saw an explosion in all kinds of messaging applications,” she explained.
The increased need for digitizing in-person services has sparked a wave of demand that’s reached all corners of the world, resulting in some of Tangible AI’s projects reaching as far as Nepal, Myanmar, and the DRC. Changa, another Tangible AI prototype chatbot, helps mothers to newborns and young children in rural areas follow the health workers’ instructions on home care for their child, as prescribed by IMCI protocol. “It’s exciting to see chatbots turning from something that only a few people used and experimented with into something that we will increasingly see as part of the toolset that organizations utilize to empower the people that they serve,” Dyshel said.
Pro-social artificial intelligence
For data scientists who are actively building AI, Lane stresses the importance of using long-term thinking. Instead of focusing on near-term profitability for consumer-driven decisions, data scientists should try and optimize algorithms for solving genuine human needs. “The problem is that a particular ‘cost function’ or ‘loss function’ is tuned for the objective of whoever is training the bot or the algorithm, the machine learning algorithm, the predictive algorithm, or whatever it is you’re trying to build with the data science,” Lane explained.
Tangible AI wants to turn its attention to “pro-social” artificial intelligence: in other words, making machines more human. For Lane, this is imperative if humanity has any shot at fixing long-term societal afflictions. “We’re giving the machines a long-term perspective to recognize human beings—their needs, wants, and desires. And companies need to recognize that focusing on those long-term needs, wants, and desires aren’t always going to create dollars in the near-term.”
Is machine learning engineering the right career for you?
Knowing machine learning and deep learning concepts is important—but not enough to get you hired. According to hiring managers, most job seekers lack the engineering skills to perform the job. This is why more than 50% of Springboard’s Machine Learning Career Track curriculum is focused on production engineering skills. In this course, you’ll design a machine learning/deep learning system, build a prototype, and deploy a running application that can be accessed via API or web service. No other bootcamp does this.
Our machine learning training will teach you linear and logistical regression, anomaly detection, cleaning, and transforming data. We’ll also teach you the most in-demand ML models and algorithms you’ll need to know to succeed. For each model, you will learn how it works conceptually first, then the applied mathematics necessary to implement it, and finally learn to test and train them.
Find out if you’re eligible for Springboard’s Machine Learning Career Track.
The post This Startup Wants to Build Artificial Intelligence With a Heart appeared first on Springboard Blog.