In late November 2018, I delivered the opening keynote at the Future of Work 2018 Conference hosted by the University of Melbourne Center for Workplace Leadership[1], where I described how to keep people at the center of the future of work discussion. After my talk (find a video of the 45 minute keynote here), I was interviewed for a short video segment to share the highlights. In preparing for the interview, I found the questions to explore a thoughtful, comprehensive look at the scope of the discussion on the future of work, and thus I wanted to share my answers.
I am interested in your reactions. Do you agree with my thoughts? Disagree? Why?
What are some examples of human-centric approaches to work within areas of high automation?
I would approach this question differently. While there are some jobs that might have a higher likelihood of automation than others, especially those contain significant amounts of routine tasks[2], the World Economic Forum has recently predicted that greater than 50% of all jobs will require significant reskilling and upskilling by 2022[3].
This means we will need to consider more human-centric approaches in most jobs, not just those with a higher proportion of tasks that could be automated.
In the research I’ve been conducting in partnership with Bill Eggers and Amrita Datar from Deloitte and Carlos Teixeira from the Harvard Kennedy School through my current part-time fellowship with the Harvard Kennedy School digitalHKS initiative, we’ve envisioned 6 government jobs of the future with a method that puts humans back at the center for both new and changing jobs. Putting the worker first, we envision future jobs to place human-machine coordination at the core, in order to improve productivity and decision-making while also increasing the well-being of the workers themselves. The trick is not to think about what can be automated, but also what can be augmented. If a human’s time could be freed up, what higher order value could be accomplished with uniquely human skills? Might we be able to deliver more value than human machine coordination than through humans alone or machines alone?
For example, Child case workers in Colorado currently spend less than 10 percent of their work time actually interacting with children and families. This impacts their ability to most effectively protect vulnerable children and coordinate services for those children. Case workers today spend a large percentage of their time on administrative tasks. Our research has envisioned a future where administrative burdens are reduced with the application of technologies to increase the time spent on higher valued work for the customer. For example some technology applications could include: case management systems powered by cognitive computing to ease scheduling and follow up; voice-based smart assistants that document case notes after an appointment while the future child aid coordinator travels between appointments; robotic process automation tools that automatically process a client’s eligibility for services on the fly; and using predictive analytics and machine learning to prevent abuse, neglect, and ill treatment. Child aid coordinators should be empowered to customize services for vulnerable children based on the child’s own unique circumstances. Spending more time with children and families is a critical part to increasing the value created by this work and by applying technology to free time from administrative tasks, coordinators are able to spend more time problem solving for their customers.
What frightens you or stirs you the most about the shift from human power to machine power?
MIT Professor David A Mindell talks about three myths of 21st-century robotics and automation in his book, “Our Robots, Ourselves” that I think are particularly relevant to shift the focus of this question:
· “The myth of linear progress, the idea that technology evolves from direct human involvement into remote presence and then to fully autonomous systems.
· The myth of replacement, the idea that machines take over human jobs one for one.
· Myth of full autonomy, the utopian idea that robots, today or in the future, can operate entirely on their own”.[4]
So, there’s a lot of fear mongering about how quickly and how extremely a “shift” to machine power will happen. Also, I think what we have to remind ourselves is that human power alone and machine power alone won’t be as great as power achieved through human-machine coordination.
To get the most value we need to design machines that have humans in the loop and empower humans to design work processes where machines are in the loop.
This is where the greatest value lies but is also where the challenge is because most organizations are not good at integrating new technologies into their business processes. One reason they struggle with this (not far from the only reason) is significant workforce resistance when new technologies are forced on them. This is where employees seeing WHY the change is important and why there is a benefit to them to re-skill is important. Another reason is because true user-centered design in the technology development and deployment process is not common practice. Organizations must embrace a user-centered technology adoption mindset if technology initiatives are to succeed in improving the performance of an organization. Co-creating a visualization for the jobs of the future (as occurs in the workshops Bill, Amrita and I have developed) can be a powerful tool to enable these actions.
How futuristic or progressive is the research in this area – are we finding out the facts and figures too late?
There is a lot of research being done in this area: but often in silos. Conversations about future of work are often segmented between communities. People building technologies focus on a technology-enabled future. Human capital and staffing professionals focus on how workforce changes are affecting the future of work. Universities and educators focus on future skills and learning mindsets as being key enablers for the future. The urgent need is to build bridges across these communities and research areas since I believe solutions will lie at the intersections of the complex topics, including, but not limited to:
· Technology development (robotics, computer and data science, engineering, etc)
· Economics and economic development policy (local, state and national)
· Education and learning
· Workforce development policy
· Technology ethics
· Labor relations
· Business and public sector management
· Human centered design/ human factors
· Architecture and build environment design
· Business strategy
· Organizational design
· Performance management
· Organizational and Individual Psychology
MIT is one university trying to get ahead of this need for interdisciplinary, action-oriented research through the “Task force on Work of the Future” announced off earlier this year but significantly more work can be done in convening these communities for active solution development.
What’s critical to do right now in ensuring the human evolves and doesn’t disappear from meaningful work?
In the near term, government, employers, employees, and educators all have a role to play to design a future that is better for humans, not worse.
Employees must not wait around for skill development to be presented to them by their employer: a mindset of life-long learning will be a key attribute for the future employee. Learning doesn’t necessarily mean traditional university higher education where costs are becoming unsustainable. Learning can happen online, through local boot camps, by doing, through micro-courses designed to be taken on-demand, etc. Employees must see that change is coming, understand the new skills that will be required for them to succeed in a changed work world, and feel empowered to take some of that learning on for themselves. Now that’s not to say that employers won’t also have an important role to play in training employees in the skills of the future, like AT&T’s $1B commitment to re-train half of its workforce for the jobs of the future. But, McKinsey estimates that globally some 500M-900M future workers may need to move to new employers and sectors to find meaningful work[5]. That transition will require some initiative on the part of the employee. Thus, we must create means for individuals to feel empowered, and encouraged to do this—and reduce fear of the future that could be paralyzing.
Educators act as a bridge between the future workforce and employers. Being an effective bridge should involve identifying the salient knowledge, skills and abilities required for the future of work and evolving curriculum and learning environments to prepare students for those futures. This requires educators to have a keen understanding of the needs of industry. Many institutes of higher education are also completely rethinking how they teach and how students learn and are making education to non-traditional students more broadly available through online learning, certifications, and MOOCs. Innovating the learning process itself, and the funding model for learning, is a critical role of educators—and a role all educators should be moving to adopt in the near term.
Employers must step out of their comfort zone and think about a future where their potential workforce could include not only be full time human employees, but contract employees, the crowd, strategic partnerships, and machine intelligence. How do you think about integrating the machine as part of the team? How does this change the role and skills for managers? How might humans and machines combined create a more valued product or service that you provide today? The challenge is to not think about the application of machines as an efficiency gain to cut overall costs by reducing human labor, but to think about how machines augment humans who can provide higher order cognitive value to the organization.
Employers should also actively engage their workforce in designing the jobs of the future: jobs that use technology to augment current work and that add new value. Current employees see problems on the front line and opportunities for additional value creation that management and senior leadership often doesn’t see. That front-line intelligence should be incorporated into designing jobs of the future. Designing a future organizational model that harnesses an agile way to leverage human-machine coordination to add value is a business imperative. Further, employers have to be obsessed with creating a learning environment in their organizations. This environment should teach existing employees how to fail fast and adapt to changing technologies and customer information regularly, but also help new employees onboard much more quickly due to the increasing fluidity of the workforce and increasing incidence of telework/ remote teams.
Governments must set economic, workforce development, training, education and technology policies that consider the interdisciplinary issues of the future of work. But their role is not only limited to policy making through laws and executive orders. They could also:
· Leverage convening power to create a cross-sector agenda
· Encourage experimentation at the state and local level as test beds for national policy
· Fund programs to re-skill employees in parts of the economy that might not be able to afford to do it themselves (like small to mid-size manufacturers)
· Consider government agency roles and responsibilities to see if policy making across different interdisciplinary areas should be better coordinated
· Encourage policy making that includes a participatory element, to ensure the voices and values of the public (not just organizational stakeholders) are considered when evaluating policy solutions
· Ensure trade-offs between technology ethics and societal values are considered in economic growth policies
· Initiate diplomatic efforts to address inequality issues in the future of work
· Etc
McKinsey[6], Bain[7], and others have also made recent recommendations on what these groups can do to prepare for the future of work.
How do we rein in the growth of inequality, as technology shifts the balance of power and opportunity in society?
New technology throughout history has been accompanied by fear of that technology. Malcolm Frank and others in their book, What to Do when Machines do Everything note: “Many seem to forget that throughout history, automation has provided a net benefit to society. In the process of automating our work and our society, through generation after generation, three positive things have repeatedly occurred:
1. A new abundance has been created; sales of products and services produced by automation — now vastly more affordable and of higher-quality —skyrocket
2. With the new abundance, overall employment rises, even when there is less labor input per unit.
3. Society gains a net benefit, with higher living standards created by newly affordable products and services.” [8]
However, this fourth work revolution does have some unique issues to consider with respect to furthering inequality. A primary issue involves the coding of biases into future decision-making structures and thus significant discussions have emerged on the ethical use of AI. This discussion can often be about how to ensure that the algorithm developers aren’t coding their own personal biases into technologies. But it can also be about the data that we are using to train our AI; if training data is biased or not representative of broader society then we could be making incorrect diagnoses faster or forgetting entire populations in the future we’re optimizing for.
For example, a 2010 report from the University of British Columbia warned against the “danger of assuming that all of [humanity’s psychology] closely matches the behaviors of 20-something college students. They cite evidence that between 2003 and 2007 undergrads made up 80 percent of study subjects in six top psychology journals, and that 96 percent of all psychology samples come from countries that make up only 12 percent of the world’s population[9]”. The researchers refer to the sample population as mostly coming from Western Educated Industrialized Rich Democratic (WEIRD) populations. So, if a future “AI therapist” were trained off this data set, the accuracy of diagnosis and the effectiveness of treatment may likely underperform with underrepresented communities in the dataset. We must be vigilant against potential bias both in our training/ research data as well as the humans coding the rules else we are likely to further encourage inequitable application of technology.
What can only humans do, and how will this change – improve, be compromised – in the future work economy?
Humans are still vastly better at problem solving, communication/ knowledge sharing, collaboration, leadership, creativity, empathy, and the ability to cope with ambiguity and uncertainty.[10]
In some jobs that conduct routine, cognitive tasks, machines can actually become smarter than humans when powered by vast amounts of data. These jobs may include call centers, auditing, insurance underwriting, and basic tax preparation. Some creative, cognitive jobs may also be replaced. For example, powered by significant amounts of viewer data Netflix can set the parameters to create new TV series and movies they know will be enjoyed by their population. As described by Jessica Brillhart, AI is also creating beautiful art.
This means humans will need to perform more sophisticated cognitive tasks in the future of work. Mohanbir Sawhney suggests a few such jobs in Forbes, like managing and designing the future human/machine teams and jobs that require social intelligence, like providing care to humans (nurses, counselors, physical therapists). Also, he notes, as technology improves the quality of life for society, humans may have more free time for leisure, which could increase employment in the entertainment and leisure industries. [11]
All previous work revolutions (agricultural and industrial) have unlocked new industries and thus new opportunities for jobs, while improving the overall quality of life of society (though there could be much vigorous debate on the nuances of this statement). If history is our guide, we would expect to see the same net change over time, though the times that society undergoes that transition can be very difficult. The relatively good news here is that the transition may be quick, occurring within one person’s lifetime, and not over the course of three lifetimes like previous revolutions. This will significantly impact a generation, but the overall transition time may be shorter. Thus, we have an urgent need to act now to minimize the negative effects on this generation, with overall progress, vigorous guard against embedded bias, and human quality of life and equity top of mind.
References:
[1]Important Note: My work on the Future of Work is conducted as an outside activity from my NASA duties. My Future of Work research is done in affiliation with the Harvard Kenney School digitialHKS initiative in collaboration with Bill Eggers and Amrita Datar from Deloitte’s Center for Government Insights. The views in my writing and speaking on this topic thus do not represent the views of NASA and are my personal views.
[2]See David Autor’s matrix describing routine versus non-routine; cognitive vs manual: https://www.stlouisfed.org/on-the-economy/2016/january/jobs-involving-routine-tasks-arent-growing
[3]Future of Jobs Survey 2018. World Economic Forum.
[4]Mindell, D. A. (2015). Our robots, ourselves: Robotics and the myths of autonomy. Pages 8-9.
[5]McKinsey Global Institute, Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation, December 2017.
[6]McKinsey Global Institute, Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation, December 2017. https://www.mckinsey.com/~/media/McKinsey/Featured%20Insights/Future%20of%20Organizations/What%20the%20future%20of%20work%20will%20mean%20for%20jobs%20skills%20and%20wages/MGI-Jobs-Lost-Jobs-Gained-Report-December-6-2017.ashx
[7]Labor 2030: The Collision of Demographics, Automation and Inequality February 07, 2018: http://www.bain.com/publications/articles/labor-2030-the-collision-of-demographics-automation-and-inequality.aspx // http://www.bain.com/Images/BAIN_REPORT_Labor_2030.pdf
[8]Frank, Malcolm. (2017) What to Do When Machines do Everything. Page 41
[9]https://www.scientificamerican.com/podcast/episode/psychology-studies-biased-toward-we-10-08-07/
[10]Adapted from http://blogs.worldbank.org/education/uniquely-human-centrality-humanism-future-workforceand https://www.fastcompany.com/40569876/these-are-the-uniquely-human-skills-that-employers-say-robots-cant-do
[11]https://www.forbes.com/sites/mohanbirsawhney/2018/03/10/as-robots-threaten-more-jobs-human-skills-will-save-us/#17a7d6fc3fce