An innovation agenda to help people win the race against the machines

Policy recommendations to get the engines of democracy firing on all cylinders.

If the country is going to have a serious conversation about innovation, unemployment and job creation, we must talk about our race against the machines. For centuries, we’ve been automating people out of jobs. Today’s combination of big data, automation and artificial intelligence, however, looks like something new, from self-driving cars to e-discovery software to “robojournalism” to financial advisors to medical diagnostics. Last year, venture capitalist Marc Andreessen wrote that “software is eating the world.”

Computers and distributed systems are now demonstrating skills in the real world that we once thought would always be the domain of human beings. “That’s just not the case any more,” said MIT research professor Andrew McAfee, in an interview earlier this year at the Strata Conference in Santa Clara, Calif.:

McAfee and his research partner, MIT economics professor Erik Brynjolfsson, remain fundamentally optimistic about the effect of the digital revolution on the world economy. But the drivers of joblessness that they explore in their book, Race Against The Machine, deserved to have had more discussion in this year’s political campaign. Given the tepid labor market recovery in the United States and a rebound that has stayed flat, the Obama administration, given an opportunity for a second term, should pull some new policy levers.

What could — or should — the new administration do? On Tuesday, I had the pleasure of speaking at a panel at the Center for Technology Innovation at the Brookings Institute to talk about what a “First 100 Days Innovation Agenda” might look like for the new administration. (Full disclosure: earlier this year, I was paid to moderate a workshop that discussed this issue and contributed to the paper on building an innovation economy that was published this week.) The event was live streamed and is available on-demand.

Below are recommendations from the paper and from professors McAfee and Brynjolfsson, followed by the suggestions I made during the forum, drawing from my conversations with people around the United States on this topic over the past two years.

Ideas from Brookings

Quoted below is the executive summary of the recommendations from the Brookings Center for Technology Innovation. The paper itself goes into more detail on each one.

  • We need better metrics for measuring worker productivity in the 21st century economy. Past approaches based on worker hours or total employees in relation to Gross Domestic Product (GDP) ignore the transformational nature of digital technology.
  • We should encourage entrepreneurship by expanding Small Business Administration credit for start ups, adding entrepreneurial skills to school curricula, and making changes in immigration policy that encourage entrepreneurs to come to America.
  • We need governments that learn to innovate and collaborate, and develop new approaches to service delivery, transparency, and participation. This includes placing more data online and employing data analytical tools, social media, mobile technology, and search results that improve decision-making.
  • We should strengthen infrastructure by investing in broadband, data centers, and mobile cell towers, and improving access to spectrum for wireless applications.
  • We should protect vital digital assets by updating the Federal Information Security Management Act and developing procedures for monitoring threats to critical infrastructure.
  • We need to improve knowledge transmission through faster adoption of digital textbooks, more widespread use of creative commons licenses for instructional materials developed with taxpayer dollars, and policy changes that speed education innovation.
  • We need to increase technology transfer and the commercialization of knowledge from universities and federal laboratories so that public and private investments translate into jobs and economic activity as well as better health, security, and well-being.
  • We should harmonize cross-border laws to promote global innovation and freedom of expression.

Recommendations from McAfee and Brynjolfsson

“Everything I’ve learned during and after Race Against the Machine has left me incredibly optimistic in all important areas except one,” wrote McAfee in an email earlier this week.

“Digital technologies are increasing our productive capacity and ability to innovate, they’re bringing good things to our lives, and they will continue to do all of the above, probably at accelerating rates. As we wrote in the book, however, as technology races ahead, it is leaving a lot of workers behind. Computers and robots are acquiring human-like skills and abilities, which means that the ‘market share’ of people in the workforce — the areas where they are superior to machines — is going down, and will continue to do so. Dealing with this trend will be one of the main challenges, if not the greatest one, that we face over the next generation.”

What to do? “The cure-all to any economic woe is economic growth,” he said, in our previous interview at Strata. Even if that’s happening, McAfee emphasized, the cohort of mid- to lower-wage knowledge workers whose jobs stand to be automated is still going to be affected.

In response, he recommended that policy makers, schools and universities rethink current educational methods and system.

“We’re turning out industrial-era workers and industrial-era skills,” said McAfee. Instead of fact-based learning methods, he suggests focusing on developing the abilities of students to do problem definition, exploration and solving, working with machines.

“Reskilling and retraining are a big part of the answer,” he said. “We’re putting the wrong skills out there. Let’s rethink that. Let’s bring government, industry, and the educational institutions together and put together a curriculum that will actually deliver valuable skills out there.”

Brynjolfsson and McAfee made a list of 19 recommendations from Race Against The Machine that are worth considering, including: measuring performance in education, extending school hours, encouraging the immigration and retention of skilled workers, teaching entrepreneurship, investing in the country’s communications and transportation infrastructure, reforming the patent system and revisiting a host of tax subsidy and regulatory policies. While not all of these recommendations could be addressed in the first 100 days, they are all worth adding to the discussion about the choices ahead of the new administration and Congress.

My suggestions

Where do I sit? To unlock more innovation in the economy, the administration should consider:

  • Making evidence-based investments in science, technology, engineering and math (STEM) education and experiential learning programs, with measurements against which teaching methods are working and where. Continue to embrace the maker movement.
  • Focusing on skills matching between millions of open jobs and coordination with industry to determine the training needed to fill them and community colleges and grad programs to create curricula that are relevant.
  • Continuing to invest in basic research and development, particularly in biotechnology, nanotechnology, materials science and alternative energy.
  • Pushing through immigration reform that enables the best and brightest entrepreneurs and researchers to come to the United States to build businesses, study and to stay here after graduation. Make the political compromise to get that engine moving quickly. Simplify visa applications, in accordance with the digital government strategy, so that innovators can build better civic interfaces to provide expedited e-services.
  • Creating more access to early seed-stage capital for startups and small businesses. That means pushing the Security and Exchange Commission to finalize the crowdfunding rules from the JOBS Act.
  • Releasing more open government data from regulators and federal agencies, particularly high-value datasets that are in demand from startups. Catalyze the growing data economy by engaging entrepreneurs and venture capitalists — and respond to their feedback about quality, availability and standards. Collaborate with states and cities on creating open standards for urban data.
  • Putting more code, research and other intellectual property created with taxpayer dollars into the public domain.
  • Using the power of government procurement to encourage small business, once Project RFPEZ is completed.
  • Releasing more spectrum and creating incentives for “last mile” broadband access to enable more participants in the innovation economy.
  • Making libraries digital hubs for communities, enabling them to provide job training, broadband access, open data curation and even as maker spaces.

Your ideas?

When I asked for feedback online before the panel, Robert Bole, director of innovation for the U.S. Broadcasting Board of Governors, commented that the administration should make “block grants of digital services and open data to states.”

Additionally, suggested Bole, the administration should “expand Code for America and Presidential Innovation Fellows. Reform Federal Acquisition Regulations (FAR), adoption of RFP-EZ (and ilk). Reform of Universal Service fund for broadband infrastructure expansion. Improve libraries and schools to provide media literacy skills. Investment of career STEM at community colleges. Aggressive restructure & auction of spectrum (with smart public use requirements).”

After she commented during the Brookings forum, I asked Gwynne Kostin what the real issues are for innovation in the federal government. Kostin, who serves as the Director of the Digital Services Innovation Center in the Office of Citizen Services & Innovative Technologies at the U.S General Services Administration, responded by putting at least some of the onus on Congress:

Imagine, for a moment, that President Obama is up late in the White House in the next month and comes across this post when he’s reading his iPad.

What change(s), policy or other action would you recommend to him in these first 100 days? Would you support any of the items from the list above? Why? Or make other ones? If you have ideas, please add them below in the comments.

tags: , , ,
  • Voltaire

    Start drafting plans for a universal living income that everyone will receive whether working or not – automation will only increase, and not everyone is going to want to be a scientist or engineer – and we certainly can’t force people to take up careers they aren’t interested in pursuing. Unemployment will keep rising as the technology progresses.

    In the future not everyone will have to a be worker bee for 40+ years of their life.

    • digiphile

      It’s always a pleasure to have Enlightenment philosophers stop by Radar.

    • Martinfsmith

      Right, but at some point the basic social contract will have to be revised. Free-market economics is two complementary systems, one to organize production and the other to determine distribution of income. The social ethic that defines this as “fair” is that your share of production is equal to the value of your contribution to production, as priced by society via the market

    • Jjoe

      I would add: strongly encourage innovations that highly reduce the costs of living. It could be an insurance against future economic troubles and a support for a low but effective universal income.

      One area best to focus here is innovations targeted at third world countries(but transferable to the first world): the relative free regulation, low cost pressures and large markets are exactly the kind of pressures you need to invent low cost stuff.

  • Doug Gibbs

    How about you redraw the shocking red line on the graphic with an axis that covers more than 20% of the whole.
    It looks like the % of population versus employment is down 5% in 17 years. The graph makes it look huge.
    The GDP line, to contrast has gone up 50%. You may have a point, but the graphic is a lie.

    • digiphile

      The graphic itself comes from Andrew McAfee, in a January 2012 post on a rebound that has stayed flat. The data behind it comes from the FRED database and the Federal Reserve. If you’re like to use it to redraw a graph, please do and then link to it from another comment.

  • I believe Paul Krugman has talked about this – or a similar graph. If you break it down the loss of employment in the US is fairly even across all sectors of employment showing an economy wide loss of demand not a mismatch of workers/skills to available jobs.

    The long term analysis maybe correct but the comparison to the recent recession is highly misleading.

  • Carl Malamud

    I’d like to throw a simple one on the stack: release nonprofit tax returns as machine-readable data (known as “Modernized E-File” or MEF format). All large nonprofits currently e-file, smaller ones have the option of e-filing. All of that data is currently released as TIFF images. If you want to take this the next step, get Congress to allow mandatory e-filing by all nonprofits: they currently prohibit mandatory e-filing for smaller organizations.

    This is a $1.5 trillion sector of our economy. In times of shrinking budgets, we will depend ever more on the nonprofit sector. Making the reporting data more usable will increase transparency, efficiency, and effectiveness in this industry just as SEC EDGAR filings do for publicly-traded corporations.

  • arsetechnica

    The Scale & Scope of the problem & response fall short by orders of magnitude. The breadth of the problem is as huge as failing to understand the agrarian/industrial demographic shift. The modern welfare state was a response to century of modern industrialization; this shift will happen on a much sharper curve. The White House has very little range to begin acknowledging the problem under the current center right-far right discourse allowed in DC. Discussion of Welfare State 2.0 isn’t currently allowed, so how is Obama’s team supposed to begin building it? Federalism is our original concept of scalability; thus planning a series pilot projects scattered from rural towns to leading cities on not a 100 day but a 3 year time frame.

    The innovation engines of tech have not yet successfully grafted on to government; an innovation czar is the current model, but a few government BDFL’s might do better. A BDFL is a bottom-up patriarch, attempts to appoint one from on high have been a sign that your once agile innovator has become a tech-zilla dinosaur.

    The WH needs a ‘forward’ team – 50 to 200 persons from a half dozen fields framing a conception of the problems building data, arguments and approaches, and a ‘scout’ team to kickstart/angel seeds across risk/yeild lines and time frames.
    Essentially, the WH doesn’t need to build a black box to fix things, but to build a black box builder. That “black box building machine” will cook Obama’s version of FDR’s alphabet soup. That’s how the 1st welfare state was built, that’s how we’ll build the 2nd.

  • We’re promoting a “Culture First” philosophy. That means we believe that changing the culture of government should be prioritized ahead of changing policy. A culture of innovation in government would not only create value within government, but also build empathy toward innovators in the private and nonprofit sectors, as well.

    Policy is important, but without antecedent cultural shifts, the impacts of those policy changes are greatly inhibited.

  • Jozef

    Good article, but I’m missing the most important topic to be addressed. What about climate change? You seem to ignore that the earth has its limits for economic growth.

  • John Scott

    Simplify to optimize. Case in point each government Agency/Department is allowed to create its own list of regulations, rules and policies around procurement, security, etc. This leads to a fragmented market of niche providers who understand how to work with/at DHS, VA, HHS, etc. leading to increases in cost, niche solutions for enterprise problems and less competition. The exec branch needs to reorganize how it works.
    1. Stop the proliferation of policies, OMB should set one and apply to all
    2. Simplify the security model
    3. release all tax-payer code as open source / public domain. We paid for it!

    Too many rules, leads to market failure and too much overhead everywhere.