Artificial intelligence has become quite the buzzword in the past year. Salesforce Ventures and Bloomberg Beta have built specific funds targeting the category. Meanwhile, exits are positive for AI startups, courtesy of major tech companies looking to bolster internal AI efforts.
But while tech startups are embracing the AI terminology in their marketing materials and VC pitch decks, the technology often claimed to be AI is actually smorgasbord of machine learning, advanced algorithms (also known as deep learning), and natural language processing—with a dose of humanity to help tape it all together.
And it’s that last bit, the human bit, that founders and investors tend to gloss over. After all, the whole point of artificial intelligence is to make our gelatinous brains unnecessary. Those brains cost money to maintain, and that’s not helpful for revenue multiples. But even mundane tasks, such as booking a hotel room, are difficult problems to accomplish with algorithms. And major tech companies are known to rely on a crew of Mechanical Turk workers to smokescreen as your AI compatriot.
However, there are signs that AI startups are growing more honest about their helping hands. And looking ahead to 2018, humanity could actually become an asset (or at least come out of the shadows) for startups claiming to be developers of artificially intelligent software.
Human Help Needed
While a tremendous number of AI startups and large tech companies go to great lengths to gloss over the human input needed to make their AI products work, one startup, Fin, swooped in at the end of the year with a notably different tact.
The AI-powered assistant, co-founded by Sam Lessin and Andrew Kortina, aims to handle any digital task at a cost of $1 per minute. And as Lessin told TechCrunch, looking to AI to replace human intelligence is a “mistake” and “the future is people helping people.”
To make this happen, Fin has 100 full-time employees that help complete tasks. It’s this human aspect that allow it to work on a dizzying array of complex requests, many of which you can see being made in real-time on the company website. But what’s most surprising is what little work Fin and company undertake to cover up the fact that humans are a critical component of the app. Its work page, for instance, gives “human” placement over “machine intelligence.” Contrast that with X.ai, which avoids discussing that humans most certainly, to the surprise of its users, share in the task of scheduling calendar appointments,
And while Fin hasn’t received VC funds, according to Crunchbase data, its honesty could lead to the rise of what some call “augmented intelligence,” or software that uses AI methods such as machine learning to simply help humans do their jobs more effectively.
It’s a term that expresses a bit more honesty in what artificial intelligence methods are actually capable of accomplishing today. The billions behind Google, Facebook, and Apple have not been able to use software to solve the most basic of requests. And if those tech giants can’t accomplish buying plane tickets automatically, it is much less likely a startup with comparably little data and engineering talent will find a software-only solution.
So while it may have been sexy for startups in 2017 to use artificial intelligence as a buzzword, technical limitations may limit how long VCs and users will believe the claim. To counter this, startups may be required to lean into their humanity rather than dismiss it as superfluous—much like Fin has chosen to do.