How Gig Economy Platforms Like Uber Will Automate Human Work

There is a class of businesses – let’s call them “Gig Economy platforms” – that uses technology to engage large pools of new labor in carrying out its work. A “platform” is something on which others can build. Uber‘s platform allows amateur drivers to compete with the taxi business. Airbnb‘s platform helps people rent their homes in competition with commercial lodging providers. Facebook‘s platform enables people to publish pictures, news and other content in ways that have significantly expanded the media landscape.

[Tweet “Like a Lego board in the sky, platform businesses make it easy to plug our work into their work.”]

Gig Economy platforms are exciting because they create opportunity. But these platform are also starting to raise questions of economic equity and, in some cases, risk automating a great many people out of work.

What Are Engagement Platforms?

Here’s a quick framework for thinking about engagement platforms:

• “First-order engagement” is a firm’s management engaging employees in the work of the firm.
• “Second-order engagement” is employees engaging customers and partners in the work of the firm.
• “Third-order engagement“ is customers and partners engaging other customers and partners in the work of the firm.

Platform businesses specialize in building software systems that enable their third-order engagement. In doing so, they don’t just mimic, but radically reinvent, traditional ways of doing business.

At a simple level, there are two business models of engagement platform businesses. Facebook, reddit, DeviantArt and BuzzFeed use a user-generated content model through which users distribute content to lots of people. eBay, Uber, Airbnb, and Amazon‘s Mechanical Turk use a marketplace model that enables users to sell to lots of people.

Goose & Golden Egg

When push comes to shove, publicly-traded corporations default to “shareholder primacy”: the idea that the primary role of business is to maximize returns for shareholders. When engagement platforms serve this idea, they ultimately extract value from stakeholders in order to concentrate it in the hands of shareholders.

The question is – for businesses that rely so heavily on external users and partners for the value they create, how sustainable is this strategy? Can we imagine alternative approaches for platform businesses that might be more mission-centric and better at ensuring the long-term health and economic viability of their stakeholders?

In Platform Cooperativism vs. the Sharing Economy, Trebor Scholz questions why stakeholders don’t build their own platforms. In his book, Who Owns the Future? Jaron Lanier asks similar questions about compensating users for their contributions to platform businesses. What would it take, for example, to build a commercially-viable, and driver-owned, competitor to Uber, or a user-owned Facebook?

These aren’t easy questions to answer, but there’s another reason why it’s very important that we try.

Gig Economy Platforms and Automation

There is a certain class of platform business that is likely to lead to large-scale automation and job loss. I’m not talking about user-generated content platforms like Facebook or marketplace platforms like eBay, since neither of these models compensates users for their labor. It’s the marketplace platforms that supply labor for the Gig Economy – platforms like Uber, Lyft, and Amazon‘s Mechanical Turk – that present the greatest problem over the long-haul.

These labor marketplace platform companies are using human labor today in order to build scale and market dominance because human labor is what works today. But machine learning and automation will ultimately reduce the labor costs of these businesses quite drastically over time. In the most extreme shareholder-centric vision, there is no labor – only a “perfect profit machine” with capital investments buying more and more automation and generating more and more profits for shareholders.

From this perspective, human labor in the Gig Economy is just a temporary solution for work that machines haven’t yet learned.

Once they’re technologically, legally, and economically able to compete with human drivers, is there any doubt a well-capitalized company like Uber will move to self-driving cars? Or that drivers for Amazon‘s new one-hour delivery service will eventually be replaced by drones or other automation?Amazon‘s Mechanical Turk helps users outsource jobs as “Human Intelligence Tasks” to people around the world right now, but it’s clearly just a matter of time before much of this work is transformed into artificial intelligence tasks. Concierge services like Magic+ and Facebook‘s “M” are both very open about their intention to move more and more of their work to artificial intelligence. Yes, these services will create near-term economic opportunities for people in picking up laundry, buying flowers and other tasks, but ultimately much, if not most, of this work will be done by artificial intelligence.

AI or IA?

In his book, Machines of Loving Grace, John Markoff writes about an ongoing tension within the computer industry that pits an artificial intelligence (AI) vision against an intelligence augmentation (IA) vision. The AI vision replaces humans with machine intelligence while the IA vision uses technology to empower humans. Markoff paints a compelling history of the computer business fluctuating over time between these poles.

As we now stand before a great wave of coming automation, this split between the AI and IA visions raises an important question for our future relationship with computing power. Do we design for eliminating ourselves or empowering ourselves?

It’s a bit of an oversimplification, but shareholder primacy does align closely with the AI focus on replacing humans. For companies coded to maximize shareholder returns, the promise of replacing human intelligence (and costs) with machine intelligence is a powerful pull. This, in turn, prompts another question: will we allow this shareholder primacy model to continue to dominate our vision for business, or will we make room for new approaches that place greater value on caring for stakeholders and serving missions that are more meaningful than simply maximizing returns for shareholders?

This decision of how we define the purpose of business turns out to be a powerful kind of “code beneath the code” shaping not just the software we build but our relationship with machines and the future of the human experience.