Learning Detours: The Hidden Impact of Driverless Cars on Education

As a general-purpose technology, AI will create cascading impacts across society. We need to think more about the interrelationships among sectors and impacts.

TL;DR

• AI will drive wage suppression and unemployment among some of society’s most vulnerable families, at least in the short-term.

• Retraining and new jobs won’t necessarily solve this problem.

• Economic insecurity will undermine education, even if AI’s benefits will improve it in narrow ways.

• Schools need to engage in a broader discussion about the societal impacts of AI, as these impacts will have consequences for their schools.

__

You are probably wondering what driverless cars have to do with teaching and learning. What’s the connection?

Well, first, let’s briefly review what Waymos are.

Waymos are driverless (driven by AIs instead of humans) “taxis/Ubers/Lyfts” that are now providing 150,000 driverless rides, totaling more than 1 million miles a week.

We can see the explosion in use here.

While the service is currently limited to 4 cities — Austin, LA, Phoenix, and San Francisco — Waymo has recently raised $5.6. billion on a $45 billion valuation to expand to more cities. This includes Atlanta, where they can be reserved through Uber.

More driverless car (Zoox) and trucking companies (Kodiak) are innovating and coming online.

It may be a while (or maybe never) before we have Level 5 driverless cars that can driver anywhere, but even current L3 driverless cars that can drive in areas that have been entirely electronically mapped will have a substantially disruptive impact in every city they are deployed.

How so? They will replace human drivers. This is already putting pressure on drivers’ wages, and Waymo wishes to capture more of this market in the future.

How will Waymos and other driverless vehicles impact wages, unemployment, and poverty?

As noted, we’ve already seen pressure on wages for drivers. As more Waymos hit the road, wages for remaining drivers will likely decline, and eventually we will not have enough rides for all remaining human drivers—certainly not a wage they can afford to drive at.

Since taxi drivers are “gig workers,” they cannot claim unemployment, and technically, they won’t have “lost their jobs” — they will just have fewer economically valuable rides to give and, eventually, no rides to give at all.

What does this mean for unemployment trends?

Debates have begun about AI's impact on unemployment. While we don’t know how things will pan out for sure, optimists make arguments in favor of the claim that AI will only have a minimal impact on jobs.

(A) AI can do many tasks, but it can’t do all of the job tasks a human does.

(B) There will be new jobs.

(C) Job training solves.

(D) Expanded AI triggered by economic growth will create new industries and new jobs.

(E) AI will “augment” workers, enabling greater productivity and related worker utility. Just learn to use AI, they say.

Let’s look at each of these.

(A) “AI drivers” can do all the tasks human drivers do, so the “tasks” argument does not apply in this context.

(B) Most taxi drivers will not be able to do any of the jobs described in B. Of course, these aren’t the only jobs, as they generally have limited education.

Of course, this doesn’t mean there aren’t any other jobs. Amazon, for example, is building new warehouses and hiring human workers who do not need advanced education. But they are already integrating robots, and as that industry expands, the number of human workers it needs will decline.

(C) Taxi driving is barely minimum wage work. Given that a driver has to pay their own vehicle expenses and usually has no option for health insurance and retirement, it’s probably less than that. Anyhow, anyone relying on this income to live is only doing it because they have to.

If taxi drivers could simply retrain for better work, they would have already done so, as almost any job would be better. And where are the free or affordable job training programs available to them now?

If job training was the solution, existing taxi drivers would be engaged in it.

(D) It’s likely that economic growth creates jobs, but what makes this industrial revolution different from previous ones is that at least many of new the jobs do not have to be done by humans. In fact, there is a race to create the first $1 billion company with only one employee.

Most taxi drivers will not be able to do the new jobs in (D).

(E) Even the drivers that know how to use AI will lose their jobs/wages, as adding a human drive to the driverless car has no value.

The Consequences

In this context, what new jobs do we imagine unemployed taxi drivers will do? In fact, most jobs that involve repetitive work are likely to be automated. Our unemployed taxi drivers are unlikely to be the financial beneficiaries of automation and unlikely to find new jobs that cannot be automated.

What is the demographic and social make-up of taxi drivers?

[This data was assembled from Perplexity and converted into these charts by Claude. While every detail of the data was not verified, the data and representations are consistent with widely reported information. I believe it is more than appropriate for a blog post].

So, generally, we are facing a world where a number of individuals who are barely getting by are facing wage suppression and eventual job loss.

In his recent book, Ray Kurzweil argued that transportation employs approximately 10% of the economy, with 5% coming directly and another 5% in industries that support drivers (rest stops, for example).

What does this mean for learning?

Last spring, I worked on an AI Guidance document for a school district with a high number of English language learners. The document is very forward-thinking and commits to providing opportunities for all students and staff to thrive in an AI World.

In this district, 96% of the students in the district are Hispanic and 87% of the students are designated as coming from low-income families. 40% of the students in the district are English language learners, with the most common languages spoken at home being Spanish, Vietnamese, and Khmer.

When any of these families start to struggle even more from job loss due to AI, combined with the fact that it will be harder for them to access job training programs in English, their children’s education will suffer.

[source]

Loss of limited income in families with limited English-speaking skills is likely to increase poverty and undermine their education.

Everyone who works in education knows the impact of poverty on learning is devastating. When children from low-income families walk into their first classroom, they're already nearly a year behind their peers academically. This isn't because they're less capable or their parents don't care—it's because poverty creates a perfect storm of challenges that directly impact learning. Their developing brains are affected by constant stress, inadequate nutrition, and limited access to the enriching experiences that more affluent families take for granted.

The daily reality for these students is harsh. Imagine trying to concentrate on algebra when you haven't eaten since yesterday's school lunch or attempting to complete homework in a crowded apartment with no quiet space to study. Many of these students miss school frequently, not because they want to, but because of unreliable transportation or the need to care for younger siblings while parents work multiple jobs. These challenges create a devastating cycle.

What link is stronger?

There is a lot written about the potential benefits of AI in education, including personalized learning paths that adapt to each student's pace and style, automated grading and feedback that gives students immediate responses to their work, intelligent tutoring systems that provide 24/7 homework help and explanations, administrative automation that frees up teachers to focus more on direct student interaction, and data-driven insights that help identify struggling students early and track learning progress more effectively.

There is debate about these, but let’s assume they are all true.

And let’s also assume that AI leads to at least some job loss (at least short-term job loss), including for the most vulnerable. This doesn’t have to be a massive societal loss, just jobs for some in particular areas. This is something that will happen.

The fundamental challenges of poverty tend to have a far more profound daily impact on learning outcomes than any potential benefits from AI educational tools. While a student might gain from personalized AI tutoring during class time, if they're chronically absent due to housing instability, struggling with trauma-induced attention issues, or distracted by hunger, the net effect could still be reduced academic achievement.

Many school administrators are currently fighting an exhausting daily battle against the symptoms of socioeconomic distress—managing escalating behavioral incidents, addressing mental health crises, coordinating with social services for homeless students, dealing with chronic attendance issues, and supporting overwhelmed teachers who are themselves struggling to manage classrooms affected by trauma and instability.

When principals are spending their mornings finding emergency mental health support for students in crisis, their afternoons handling serious behavioral incidents, and their evenings working with families facing eviction, the notion of developing comprehensive AI integration strategies becomes almost absurd. The bandwidth simply doesn't exist to take on major technological initiatives, including what some call “transformational initiatives” that are supposed to restructure the entire sector (which also makes people nervous as they fear their jobs will be lost in the “transformation”), when basic student needs and safety concerns demand immediate and constant attention. While AI vendors and education technologists promote visions of digitally enhanced learning environments, the reality is that school leaders are consumed with ensuring students have stable housing, adequate food, and basic emotional support—fundamental needs that must be met before any meaningful learning can occur, with or without advanced technology.

Time for a Societal Discussion

In the first presentation I ever gave on AI and education back in April 2023, I explained that while there was a debate about it, schools should start to prepare for a world where families experience job loss.

While the debate about the net impact of AI on employment and quality of life continues, there is no doubt that it will take some people’s jobs, at least in the short-term. In instances where families are already living at the edge or below the poverty line, this will sting especially hard. At a minimum, this will impact students’ learning.

Recently, the former Director of AGI safety at OpenAI, stated that “the world is also not ready” for AGI. While debates about AGI (Artificial General intelligence)—machines that can match or exceed human-level cognition across all domains - are interesting, we're already struggling to manage the societal impact of narrow AI systems that excel at specific tasks such as driving taxis. We haven't developed robust social support systems, job transition programs, or economic safety nets for workers who could be displaced by AI capabilities we have right now. Mainstream political discourse related to these programs includes claims that their advocates are “socialists,” completely ignoring that I can’t think of a single leader in AI who doesn’t think these types of programs will be needed, at least in the transition to a better, even utopian, AI World.

While a single instance, the taxi rideshare driver example is particularly telling—we already have companies testing and deploying autonomous vehicles while many of their current human drivers are struggling to earn living wages and access basic benefits. The fact that we're failing to address these immediate challenges with narrow AI suggests we are woefully unprepared for the broader disruptions that more advanced AI systems could bring.

The urgency for a comprehensive societal dialogue on AI’s role in our future is unmistakable. AI is not just another innovation; it’s a general-purpose technology that will influence nearly every facet of society, from healthcare and law to transportation and finance. Because all these areas are interconnected with child development and education, the impact of AI cannot be viewed in isolation. If we focus solely on the “Edtech” component—the tools and technologies in classrooms—we miss the broader picture of how AI shapes the world students are preparing to enter.

AI-driven shifts in the workforce, economic stability, social services, and even civic engagement will directly affect students’ lives, experiences, and aspirations. Schools need to start thinking about these things and how they will address these issues with students.

Conclusion

AI's progress will profoundly affect work, learning, and well-being. If we fail to address these shifts, individuals and families will be left vulnerable, facing challenges like job loss and disruptions in educational systems without sufficient support. Our debates must go beyond hypothetical scenarios to consider the concrete actions we can take today to build resilience against the disruptions AI may bring tomorrow. Just as we need a roadmap for AGI, we urgently need clear pathways and robust frameworks for navigating AI's impact on our workforce, educational institutions, and social safety nets.

If we are going to actually “transform” education, we will need to do a better job of transforming the world. This includes preparing for the “Human-Led” AI transformation we are bringing about.