Blog: The Future of Work Is Bright… If AI-Produced Wealth Can Be Shared
Artificial Intelligence (AI) has the potential to massively increase economic output. Machine learning, industrial robotics, computer vision, and virtual assistants are all forms of AI that will increase productivity and efficiency. Accenture projects that AI could double global economic growth rates by 2035, and McKinsey projects that AI will generate an additional $13 trillion of GDP by 2030.
While those headline statistics sound exciting, they don’t tell the whole story. Buried in these projections are millions of workers who stand to lose their jobs over the next 5–10 years due to automation. The University of Oxford’s seminal study on automation and the future of work estimated that up to 47% of work activities have the potential to be automated by 2030. Today, I’ll be examining how AI is drastically changing what we work on, how we train for it, when and where we work, and what AI means for collective prosperity.
What We Work On
McKinsey published a useful infographic on the technical potential of automation across a variety of work activities and industries. Unsurprisingly, they found that the most automatable activities are predictive and physical in nature — including food service, manufacturing, and transportation.
Unpredictable physical labor, found in construction, human care services, and agriculture — is much more challenging to automate. It’s surprisingly difficult for a machine to unload a dishwasher and organize a kitchen cabinet or spoon-feed an elderly hospital patient.
Machines aren’t just replacing physical labor. McKinsey found that certain cognitive activities — such as data processing and collection — were also highly automatable. These activities are performed in massive industries including financial services, retail, warehousing and administration.
As AI and machines are able to perform these work activities faster, better and cheaper than humans, the question is, where will workers go? McKinsey reported that the least automatable activities include managing and interacting with people, creativity, and decision-making. Thus, many workers will step into managerial roles, overseeing and coordinating fleets of industrial robots. Others might become quality control inspectors and repair technicians. Others still will transition into software developers and solution architects. Many will be replaced by machines altogether, and will need to transfer their skills to growing industries, such as solar and wind energy, construction, care work, and, of course, technology.
How We Train For Work
Long gone are the days when we study one discipline in college and commit to one field in the workforce. Given the potential for AI to transform entire industries as described above, workers will need to become adaptable, lifelong learners. Education technology tools — such as massively open online courses (e.g., Coursera and EdX) — provide low-cost access to thousands of university courses in a flexible, self-paced learning environment. Microcourses are short bursts of just-in-time education on essential skills, often scenario-based. Extended Reality (XR) and immersive technologies can also be used to accelerate training and upskilling in simulated environments.
How we pay for skill training is also shifting. Income share agreements (ISAs) are increasingly being offered by universities (e.g., Purdue University) and private investors to provide funding for postsecondary education in exchange for a percentage of a student’s future earnings. While some ISAs bear onerous repayment terms, worker-friendly alternatives have arisen. At Lambda School, a venture-backed ISA company, students learn for free and don’t pay until they earn at least $50,000, and owe nothing if they don’t. Lambda’s programs include full stack web development, data science, and UX design.
When And Where We Work
A recent Gallup poll found that 57 million U.S. workers (36%) are part of the “gig” economy. Gig work refers to short-term contracts or freelance work as opposed to permanent jobs. Innovative business models have arisen to support these workers, such as Upwork, a marketplace for freelance workers, and, of course, the abundance of apps for ride-sharing, food delivery, home repair services, childcare, etc.
A 2019 study by the International Workplace Group found that 50% of business workers globally work outside one of their main office locations half the week or more. The study surveyed 15,000 professionals across 80 countries. Four out of five respondents reported that, given two similar job offers, they would turn down the one that didn’t offer them flexible working. More than 8 million Americans work from home (myself included). New business models have also emerged for remote conferencing and project management, including Join.me, Trello, Slack, Teambook, Dropbox, and Asana and just to name a few.
As work is increasingly decoupled from the workplace, so too are compensation structures. Workers are given more autonomy, yet the tradeoff is less predictability over income, work stability and benefits.
What AI Means For Collective Prosperity
The future of work is marked by two opposing forces: (1) the massive economic potential of AI, and (2) the potential for millions of people to be displaced from the workforce. A potentially dystopian outcome of these forces could be the concentration of wealth and power among the owners of AI, which, today, is dominated by a handful of companies, including Google, Baidu, Amazon, Apple, IBM, and NVIDIA, to name a few. Stephen Hawking said it best in his last Reddit post before passing away:
To avoid massive wealth inequality, it is imperative we find ways for workers and the general public to participate in AI-generated profits, whether it be through stock ownership, worker cooperatives, or creative redistribution structures. The Future of Humanity Institute at Cambridge University is currently exploring a “windfall clause” framework, whereby tech corporations agree to commit excess profits above a very high threshold (i.e., $1 trillion) to a public good fund. The framework is still in early development, but I’m eager to see how it evolves over time.
I’m passionate about the future of work because it means that more time and resources — both human and machine — can be unlocked to solve the world’s biggest problems. As AI and automation take on more and more discrete tasks, more human time and energy will be freed up for creative problem-solving. We’ll be better equipped to improve the quality of and access to food, clean water and energy, housing, healthcare, transportation, and education for more people around the globe.
At the end of May 2019, Todd Terrazas (AI LA Founder & President) and I are headed to the AI for Good Summit at the United Nations, where leaders in AI, government, media, and NGO’s will explore how AI can help advance progress on the UN’s Sustainable Development Goals (SDGs). I’ll be particularly interested to see what solutions arise for the “Decent Work & Economic Growth” and “No Poverty” SDGs.
This conversation is only the beginning. This Thursday, May 16th, I’ll be moderating AI LA’s Future of Work panel at Google Venice. We have assembled a diverse powerhouse of future of work thought-leaders, including Kian Gohar (Founder & CEO of Geolab), Kyle Jackson (Founder & CEO of Talespin), Vivienne Ming (CEO of Socos Labs), and Joe Rogers (Founder & CEO of WorkDone.ai).
Let’s create a world in which the future of work means more opportunity and prosperity for all.