Does using AI give you more time?
Today, I explore some thought-provoking research on AI adoption and its impacts on workers before turning to some interesting reads that I believe are worth your time. I hope you enjoy it!
Thanks for reading Orbit Policy's Deep Dives! Subscribe for free to receive new posts and support my work.
Subscribed
More AI Use Leads to Longer Working Hours and Less Leisure Time
Photo by Jeff Isaak on Unsplash
I’ve previously argued that if we are to get much-needed productivity benefits from AI, we need to focus on a worker-centric approach to adoption. As part of that, I’ve argued we need to emphasize and support AI adoption that augments workers rather than automates their tasks.
A collaborative and inclusive approach to integrated AI in work is needed to give employees ownership over the process and the training they need to understand the opportunities and act as champions. This is especially needed due to the well-known risks of AI amplifying biases and the significant potential for AI to erode job quality.
You can check out the full post here:
Empower Canada’s Workers to Deliver AI's Productivity Promise
In light of that, this recent NBER paper by academics from Emory, Auburn and Fordham Universities is especially interesting and challenging.
In it, they have used individual-level time diary data spanning 2004-2023 from the American Time Use Survey to understand how AI reshapes work and its impacts on work time, contracting efficiency and the distribution of productivity gains. They probe the impacts on workers with higher exposure to generative AI after Chat-GPT’s release, along with the exposure of remote workers to AI surveillance technologies during the pandemic.
Their findings are pretty dire:
The combined results suggest that while AI-driven productivity gains promise greater efficiency, they have resulted in longer working hours and lower employee satisfaction, especially in competitive markets and for occupations with higher AI complementarity, challenging the conventional expectation that technology frees humans from prolonged workdays.
Breaking that down, the authors identify several key trends.
First, AI use increases work hours, undermining the expectation that AI should enable people to finish work faster. Part of this relates to labour market dynamics, where “AI-augmented productivity motivates workers to extend their working hours” and that “occupational complementarity to AI is positively correlated with wages, which are expected to rise to incentivize greater effort.”
In theory, that sounds great. You’re more productive, so you work more to maximize earnings. The reality is less positive.
The authors find that “despite higher compensation, greater AI exposure is associated with lower general welfare of workers” and that “the extension of the working day is more pronounced when the labor market is competitive, reducing workers’ bargaining power to extract rents from technology-enabled productivity gains; or when the product market is competitive, leading to most of the rents being passed on to consumers, leaving little for firms to share with workers.” When it comes to AI monitoring of workers, they conclude that remote workers exposed to AI surveillance work longer hours post-pandemic.
All of this is concerning. Canadians are already below average in the OECD for time devoted to leisure and personal care, ranking 30th out of 41. Widespread adoption of AI could lead to further eroding of that time rather than opening up more time for workers as promised. Leisure time is also unequally shared between genders, with women taking on a far higher portion of unpaid work, so erosion of leisure time also has further inequitable outcomes.
It is also important for policymakers to consider this and think about technology and innovation more critically. As the paper’s authors conclude, “To achieve a world where humans work less and enjoy greater well-being, deliberate policy interventions, equitable distribution of productivity gains, and cultural shifts prioritizing leisure and quality of life are essential.” While I made the case in the piece linked at the top that a more collaborative model of AI adoption is needed to ensure benefits, this paper demonstrates the need to take an even wider lens.
I believe the findings also help reiterate something I often argue: innovation is neutral, and there is no guarantee it will lead to positive outcomes unless we actively work to shape those outcomes. The dominant paradigm of the past few decades, which emphasizes deregulation, a smaller role for government, and unfettered innovation, has not led to broad-based prosperity. We need new paradigms that truly focus on outcomes. As Neil Lee argued in his book Innovation for the Masses, “The true purpose of technological leadership is not to win Nobel prizes or develop the most disruptive technology but to increase living standards. Policy for innovation should aim to create good jobs and ensure prosperity is widely shared. Innovation itself is vital but only half the answer.”
We have to think about AI and innovation holistically, ensuring a coherent and consistent response that is focused on the outcomes for real people.
Reading Recommendations
President Trump’s 2025 Trade Policy Agenda - As you are no doubt aware, the United States has imposed tariffs on Canada as of today. Yesterday, USTR also released its trade policy agenda, and as Danielle Goldfarb has argued, “It is absolute. There is no nuance. It tells us exactly where Trump stands, both in tone and substance.” As she also says, it is worth reading.
It makes the case that the decline in US manufacturing jobs and growing trade deficits are “the product of a withering, decades-long assault by globalist elites who have pursued policies—including trade policies—with the aim of enriching themselves at the expense of the working people of the United States.”
Fortunately for the US, in their reading at least, “President Trump alone recognized the role that trade policy has played in creating these challenges and how trade policy can fix them.” To move towards a new “production economy,” “the United States will take action to create the leverage needed to rebalance our trading relations and to re-shore production, including, but not limited to, through the use of tariffs.” This is the quiet part out loud - it is all about leverage. That is leverage on other states, and it is leverage on foreign firms, trying to get them to move to the US, as Shopify seems to be setting itself up to do. There is no rational response or deal to be made that is a logical win-win for another state and the US. Instead, there will be the continuous ratcheting up of asks to create leverage for “the most extraordinary nation the world has ever known.” Sigh.
America’s Reckoning with Chaos and Power - Also on the US is this piece by CIGI Senior Gellow Rafal Rohozinski. In it, Rohozinski compares the actions of Trump and Musk in seizing control over the bureaucratic structure of the US as an echo of the consolidation of Bolshevik power in revolutionary Russia. He rightly points out that “Change was needed — not just in the United States, but across the Group of Seven and Group of Twenty. The world has grown more complex, while institutions have struggled to evolve, clinging to outdated models, mortgaging the future to sustain the present.” But the need for change alone is not a justification for acting without thinking. As Rohozinski argues, “here’s the problem: the transformation currently unfolding in the United States is almost entirely inward-looking, ignoring the question that matters most: What comes next?”
For Rohozinski, “The danger of tearing down institutions without a plan for rebuilding is that chaos fills the void.” For me, that serves as a strong rebuke to commentators like Sean Speer, who minimize the harm occurring in the US and emphasize the need for applying a tech mindset to public administration. Yes, much about government is deeply broken. However, we need governments and their services. They can’t be allowed to fail because of the severe harm to real people that would result. Going in like a wrecking ball is not the answer. We need a plan for rebuilding upfront.
DEI Rollback Is The “Wrong Direction for Canada,” Open Letter Says - Finally, a more positive piece. I’ve been pretty critical of some of the cowardly leadership being shown when it comes to attacks on DEI, arguing that we need to ”point out the hypocrisy and economic illiteracy of those who criticize and try to dismantle DEI efforts.” In light of that, it is welcome to see that a large section of Canada’s tech community has signed on to an open letter that strongly speaks for the importance of DEI. It also argues:
We will not allow business leaders, many of whom are unelected, unknown, and unaccountable, to dictate the future of this country. Businesses should not run the government. The government should not be beholden to business interests that prioritize profit over people.
As mentioned in the Betakit article I linked, this appears to reference the tech-bro organization Build Canada, recently established by a group of predominantly white and male tech executives and now churning out highly questionable and clearly self-serving policy memos.
Whether those companies signed on to the open letter actually embody the values they espouse and whether they are genuinely working towards the kind of systemic work necessary to build a truly equitable society and economy is another matter. However, I find it very welcome that there are business leaders willing to even put their voices out there to counter anti-DEI narratives and counter the impression that a small group of executives can set the agenda for a much larger and far more diverse tech sector.
