Metrics, not just technology, will determine who benefits from AI productivity gains

This post technology-will-determine-who-gains-from-ais-productivity-gains/” target=”_blank” rel=”noreferrer noopener”>Originally appeared on the Christensen Institute blog. and is republished here with permission.

Key points:

At the Christensen Institute, we have long argued that when a technological innovation emerges, the path it follows is rarely about the technology itself, but rather about the model that surrounds it.

That distinction will be critical to understanding the new avenues that generative ai (GenAI) can open in our schools and post-secondary institutions, particularly with regard to the dramatic efficiencies ai can generate.

GenAI is already demonstrating an immense impact in expanding human productivity. In other service industries, such as consulting, researchers have found that ai can increase the productivity of skilled workers. ai-can-boost-highly-skilled-workers-productivity” target=”_blank” rel=”noreferrer noopener”>up to 40%Some initial estimates suggest that teachers using ai could save 20 to 30% of their time currently devoted to administrative tasks.

This means that while there are almost infinite possibilities for how ai will shape what and how students learn and their creative pursuits, in the short term, much of ai will be about making the existing system more productive.

Efficiency gains are welcome developments in an education system plagued by resource constraints, but not all efficiency gains are created equal. Organizational norms and policy incentives will determine how newly acquired time and resources are reallocated and absorbed back into the system. Institutions, educators, and students stand to gain—and potentially lose—in markedly different ways.

As the market begins to chase the many efficiencies that Gen ai promises, here are three trends to watch:

1. Educator capacity: Free up time to connect…or give space to breathe?

Whether it’s teachers who spend up to 40% of their time planning lessons and maintaining student records or some counselors who spend more than a third of their time scheduling courses and administering academic tests, Gen ai has clear potential to free up valuable educators and staff time.

ai advocates sincerely hope that educators will spend their newfound time connecting with their students. Unfortunately, However attractive the proposition may be, it is based on the ai-can-make-schools-more-human-if-schools-prioritize-relationship-metrics” target=”_blank” rel=”noreferrer noopener”>Wrong assumption that schools are designed to optimize connection in the first place. While most educators would wholeheartedly agree that relationships matter, schools rarely measure students’ connections (with educators, peers, or community members) with regularity or rigor.

So what could happen with free time? It could start by making the jobs of educators and counselors much more sustainable. That's a very good thing. Most educators work with overwhelming workloads and mediocre salaries, leaving many educators working unpaid overtime and sometimes take on second jobs to make ends meet. If ai can offset some of that burden, it could mitigate burnout, increase retention and make teaching a more attractive profession.

Early data suggests this is already happening among frequent ai users. Aaron Cuny of ai-for-equity.org/” target=”_blank” rel=”noreferrer noopener”>ai for equity Has collected ai/player/-NxE-aQI_N1boWJelald/Slide-for-Julia” target=”_blank” rel=”noreferrer noopener”>data of staff at six charter school management organizations across the country. The data shows that a whopping 84% of those who used ai on a daily or weekly basis were “more excited about working in the continuing education sector because of ai” (compared to 52% of all respondents).

The takeaway: Gen ai has the potential to make educators’ jobs sustainable; but without new priorities and metrics, hoped for upsides like building connections are unlikely to unfold at scale.

2. Supporting students: Fixing broken systems… or maintaining them?

ai is also beginning to bring efficiencies to the Wild West of “student support.” This is especially true in higher education, where universities are scrambling to support students right up to the last minute. 40% of students who drop out of school, taking their tuition money with them, before earning a degree.

The barriers to college completion reflect the complexity of the higher education system itself. Whether it’s keeping up with financial aid, finding housing, or registering for classes, ai-powered chatbots are streamlining the extremely complex system of checklists and departments that students must navigate to stay afloat in college. In other words, ai offers a compelling alternative solution in a system that is far from student-centric.

The clearest example of this predates generative ai, when universities turned to text-based chatbots to support college persistence. Some of these models, such as Georgia State University's much-lauded partnership with Mainstay, have published double-digit gains in student enrollment and persistence.

GSU is an example in this area, not only because it measured up to benchmark RCT evidence, but also because of the organization’s commitment to focusing on student success, not just revenue. Case in point: a portion of the revenue earned from retaining students has been reinvested in recruitment. further, No fewer advisors. In other words, what might seem like pure efficiency is actually driving greater investment in student support structures.

I suspect other campuses, especially those struggling financially, may not share GSU's calculation. That raises a broader question: Are ai-enabled student support robots subsidizing a broken higher education business model or helping universities adapt their systems to be more student-centric?

The takeaway: The most promising ai-enabled student support models will use technology to better understand how to streamline their enterprise and then make changes to drastically ease navigation hurdles. But if ai is enlisted as a pure-play efficiency innovation in the traditional system, we’re unlikely to see shifts in the underlying structures that make college completion a gamble.

3. Social connectivity: Reducing costs… or making us lose connections?

Making teaching more sustainable or universities more accessible are noble goals. While it may not mean a total reinvention of our education system, the efficiencies offered by ai could make the system work much better for many more teachers and students.

That said, there’s a bigger picture to consider as ai becomes part of the operating system of education, particularly with applications geared toward families and students. If it’s supported by the metrics of current ai tools on the market, relationships risk getting lost in the mix.

Because the education market is primarily focused on metrics related to learning and achievement, does not tend to demand tools that build prosocial relationships and behaviors. That means The more common ai peers, coaches, and anthropomorphized robots become in learning and support models, the more fragile students’ social connections may become. In turn, social networks that lead to lasting support and career opportunities may disappear.

As the ai ​​generation becomes more sophisticated and “personalized,” we’re going to start walking a tightrope between productivity gained and potential connection lost. That raises questions I hear few leaders in education and ai circles ask themselves: When is an ai companion a helpful “co-pilot,” and when is it undermining the time you spend building authentic connections that support your goals? When is it a helpful “assistant” that expands human potential, and when is it eroding your capacity for empathy? When is it a highly personalized “coach” that democratizes support, and when is it reducing the number of people who know you and are willing to bet on you?

The takeaway: The threats ai poses to student connection aren’t going to appear overnight. But in the long term, if productivity is at the core of most policies and revenue models that guide education, sacrificing human connection will become the cost of doing business. 

Conversations about ai and success metrics must go hand in hand

I described these possible futures as either/or options. Many readers probably hope that with the right tools and policies, ai can deliver a both and path – both to free up educators’ time and deepen connections; both to fix the current system and ultimately transform it; both to unlock individual productivity and foster diverse connections.

While I admire that optimism, let’s not forget that a new set of student-focused metrics will need to emerge to guide that growth. In investment parlance, we’ll need to see a portfolio approach to the market, investing in tools that appeal to the incentives of existing systems to seek efficiency while incubating tools that target our larger ambitions for schools and students.