How to approach the second machine age
HOME | NEWSROOM | ARTICLES

How to approach the second machine age

6 ways to get ready for the digital future.

By Kara Baskin  |  March 6, 2018

robot-typing

Why It Matters

There’s still time. Two MIT researchers explain what leaders should do to prepare their firms for a digital future.

Technology is about to transform the economy much as it did a century ago, when electricity transformed manufacturing. Now businesses need to reshape their organizations and attitudes to thrive in this second machine age, integrating three core areas: minds and machines, products and platforms, and the core (centralized institutions) and crowd.

“Technology is moving fast, but human institutions and organizations aren’t keeping up,” MIT Sloan professor Erik Brynjolfsson said in a Feb. 22 talk for the LinkedIn Speaker Series. Bryjolfsson and MIT Sloan principal research scientist Andrew McAfee discussed their latest book, “Machine, Platform, Crowd: Harnessing Our Digital Future.” The talk was led by LinkedIn co-founder Reid Hoffman.

Here’s what Brynjolfsson and McAfee said leaders should keep in mind as the second machine age unfolds. Watch the full discussion below.

 

Human judgment can be flawed. “Most organizations are not nearly data-intensive enough or technologically intensive enough. Too many decisions are made by human judgment. People are overconfident in their own judgment,” Brynjolfsson said.

He pointed to academic tenure: Professors are granted tenure based on a committee. With sophisticated analytics —  “Moneyball for professors,” Brynjolfsson said — it was possible to predict who should get tenure based on citations and publications. Comparing a data model’s list to an actual tenure list, there was a 70 to 75 percent overlap. But professors who were on the model’s list and not on the actual list went on to conduct higher-impact research.

“If you wanted people who had great research, you would have been better off going with the model,” he said.

Companies need to embrace a culture that “lets evidence speak.” McAfee referred to the danger of “HIPO” culture, referring to the Highest Paid Person’s Opinion.

“Leaders still think they need to be the gatekeepers of judgment, of ideas,” he said. Innovative companies reject this approach in favor of an evidence-driven culture.

“It’s not so much the tech; it’s the culture of how you make decisions that needs to change,” Brynjolfsson said.

“Tech Lash” exists, and tech companies need to get ahead of it. “Tech is doing amazing things for society. It’s one of the best things that can change so many things: health, poverty, and so forth,” Brynjolfsson said.

“It’s the only free lunch economists believe in,” McAfee added.

But there is a “tech lash,” they said, and tech companies have been remiss in not addressing it. Cyber-risk and vulnerability, machines doing low-wage labor, privacy abuse issues, algorithmic bias when machines make decisions based on flawed human decisions, and the proliferation of fake news has drawn negative tech attention.

“Tech companies need to take the lead and get in front of this. If they don’t, the Republican Party, the Democratic Party, the media, and lots of other people will set the agenda in ways that are not friendly to tech companies or the right kind of solutions for society,” Brynjolfsson said.

Companies should tap into the power of crowd interface. “If you have an objective benchmark and a tough problem, give it to the crowd,” McAfee said. “Pose a geeky challenge, activate a crowd, and you will get an improvement.”

“The core, people in your organization, they tend to be similar to each other. They have a certain expertise. … There’s diminishing returns doing more of the same things,” Brynjolfsson said. Using technology to solve problems allows a diverse workforce to weigh in, using sites like Kaggle, which runs data prediction competitions.

Workers probably won’t lose jobs to machines. In a question-and-answer session, the authors discussed reshaping skills for workers who could be displaced by machines.

While machines can perform certain jobs stunningly well — reading radiology reports, for example — certain jobs will always favor humans, they said.

“Almost no occupations are entirely affected by automation,” Brynjolfsson said. “As amazing as machines are, there are still so many things that only humans can do well and no shortage of problems that can only be solved by humans.”

However, the labor force participation rate is a postwar low; pairing workers with the right jobs is the issue.

“There is no shortage of work now and for a decent time into the future. We need to tweak some things in the way we match people to jobs and spend public money,” McAfee said, such as on infrastructure.

Productivity will eventually increase. Despite rapid technological changes, the rate of occupational change and productivity growth is slow. Why? Technology’s full impact has yet to be felt, they said.

“Machine learning is general purpose,” Brynjolfsson explained. “It affects lots of products and services, much like electricity and steam engines. In each case, there was 20 to 30 years between development and a significant productivity boost.”