We often overestimate the potential for AI because it’s easy to imagine computers as smart as people. Science fiction is full of them. But it’s much harder to create such machines than to imagine them.
All of today’s most advanced AI programs are only capable of specialised intelligence —doing particular tasks like recognising faces, playing Jeopardy, or driving cars. But any normal human five-year old has far more general intelligence — the ability to learn and do many different tasks — than even the most advanced computers today. Experts on average predict that human-level artificial general intelligence is about 20 years in the future, but that’s what they’ve been predicting for the last 60 years.
On the other hand, we often underestimate the potential for using computers to provide hyperconnectivity — connecting people to other people (and machines) at massive scales and in rich new ways. In fact, it’s probably easier to create massively connected groups of people and computers (like the Internet and social networks) than to imagine what these ‘superminds’ will actually do.
Superminds – such as hierarchies, markets and communities – are composed of people and computers doing things together that neither can do alone. For example, superminds use machines to do complex calculations but people to decide which programmes to run in the first place and what to do when things go wrong.
When things go wrong…
In every past case where technology destroyed jobs, markets eventually created even more new jobs. For instance, in 1900, 41% of the US workforce was employed in agriculture. By 2000, it was only 2%. But it would have been very hard in 1900 to even imagine many of today’s jobs.
Is it possible that this time will be different — that the jobs eliminated by artificial intelligence will never be replaced? Yes, it’s theoretically possible. But I think the burden of proof is extremely high for anyone who argues that the outcome this time will be different from all the other times technologies have eliminated jobs throughout history.
For instance, the printing press destroyed jobs for scribes but created many new jobs in the publishing industry. In the same way, Google’s search service reduced some of the needs for reference librarians, but it created new jobs in the online search industry, including software developers, advertising salespeople and search-engine optimisation specialists.
In the legal world, if machines can do the routine parts of legal research more cheaply than law-firm associates do today, then we will probably do more legal research, and even more human attorneys may need to apply general intelligence to decide how to use the results. Similar possibilities exist in almost all industries from generating parts of corporate strategic plans, to making predictions about geopolitical or business events, to writing pieces of text that require general intelligence or interpersonal skills.
But even if enough new jobs are eventually created to replace the jobs that are destroyed, that doesn’t mean there’s nothing to worry about. We still need to help workers who lose their jobs but aren’t able to do the new ones.
One obvious possibility is to retrain these workers. But who will pay for this retraining? An intriguing option is for worker associations such as unions or professional societies to offer a kind of unemployment insurance to their members. As a member, you would pay a certain portion of your income in the good times, in return for a guaranteed minimum income in the bad times. Then it would be in the association’s interest to help you learn new, marketable skills if your old job were eliminated.
For instance, if driverless cars took away your job as an Uber driver, your association might pay for you to learn basic web development. And since these worker associations would also be communities, they could apply social pressure if they thought you just weren’t trying hard enough to find a new job.
There are also obvious ways that governments could help with this problem by funding retraining programs, providing tax incentives for job creation, and if necessary, supporting displaced workers with unemployment insurance or universal basic income.
In other words, the most important superminds that already exist in our society — such as markets, governments, and communities — can do more than many people realise to help manage our transition to a world where new human-computer superminds do things we can’t even imagine today.
Thomas W. Malone is the Patrick J. McGovern Professor of Management, a professor of information technology, and a professor of work and organizational studies at the MIT Sloan School of Management. He is also the founding director of the MIT Center for Collective Intelligence and the author of the book, Superminds, upon which this essay is based.
Image credit: Dim Dimich/Shutterstock