You have /5 articles left.
Sign up for a free account or log in.
Three things are going to happen in higher ed in the next few years:
- Artificial intelligence will fundamentally change the business of universities.
- A new set of companies will develop, pivot and evolve to partner with colleges and universities to deploy AI.
- The new university/company AI partnership model will resemble the current online program management model, in which the AI company de-risks the AI investment with up-front capital and resources, and is paid by a revenue share.
A fourth prediction would be that the institutional AI-enabler business will end up dwarfing the OPM industry in terms of revenues and impact. Not every college wants to have online programs, and many can build and market online programs themselves. Every institution will need to leverage AI. And no school will be able to develop the sort of internal competencies that shifting business operations to an AI footing will demand.
Why do I think that these three (or four) predictions will materialize?
The answer has almost nothing to do with ed tech, which almost always seems to overpromise and underdeliver. Instead, the reason that AI will change everything in higher ed -- build the next billion-dollar industry -- can be found in a 2,500-year-old board game.
Anyone in charge of anything having to do with the business of running a university should have recognized an inflection point in their jobs the day that Google’s DeepMind AlphaGo beat the world’s No. 1-ranked go player, Ke Jie, in a three-game match. (The games were played in May 2017).
The big deal about the AlphaGo victory is the way in which the DeepMind technology works. AlphaGo was not programmed to play go, but rather the program learned to play the game. This machine learning relied on an artificial neural network to turn data on previous go games into new game-playing strategies. Later versions of AlphaGo were able to achieve domination in the game not through using prior games, but by building up strategies by playing the game against itself.
It may seem a stretch to imagine that a go-playing computer will change how higher ed is run. Skepticism is warranted. Go has nothing to do with higher ed.
The leap we all need to make is to understand how narrow AI technologies, made possible by advances in both processing and algorithms, will change how humans approach certain types of problems. The neural networks that underpin tomorrow's AI platforms will enable technology-assisted decision making that will be orders of magnitude more accurate than what we can achieve today.
Some smart people -- some of them in higher ed -- have been spending lots of energy worrying about the impact of the coming AI revolution. Higher ed people have worried about what will happen when robots take all the jobs. Or even worse, when the AI decides that humans are no longer worth the bother, and the technology decides to eliminate us all.
A better way to think about AI is to figure out all the business problems that your university faces that are like playing go.
What specific (narrow) challenges does your school face that require you to take some action based on probabilities?
What are examples of decisions and actions that people take, and which you have data on, that impact the economic viability of your institution?
Some examples that come quickly to mind are recruitment and retention. People decide whether or not to apply to your school and then to accept a place should they be accepted. Once they are enrolled, there is a nonzero probability that they will not persist to a degree.
Our understanding of why some potential students become actual students, and then why some leave before finishing, rests today somewhere between art and science. We use data as best we can to make informed choices of how we market, interact with and support our potential and existing students.
Artificial intelligence will allow the people on campus in charge of recruiting and retention -- and many other aspects such as alumni giving -- to make better decisions. The AI platforms will ingest data from many sources to develop recommendations. The more data made available to the AI platforms, the better the suggestions will be.
Like with online learning, the challenge of campus AI will not be the technology, but the implementation. The technology platforms that rely on neural networks to make recommendations around whom to market to and how, and which students should get extra supports, will ultimately be commodified. Just as the learning management systems and synchronous meeting tools that online education depends on are now ubiquitous, AI systems that ingest data to make algorithm-based recommendations will be universally available.
What will be hard is getting all the right data into the AI system and then making meaningful changes in strategies and actions based on what the AI systems recommend.
The data that will power the campus AI systems of the future will be much broader and deeper than the information currently available from campus data warehouses. The data will be drawn from an enormous range of public and proprietary sources, some of which the college or university controls and some to which institutions are currently blind.
The challenge of how to use AI to support institutional business decisions will be large enough that most schools will not be able to take advantage of the opportunity. The payoff for incorporating AI into the regular operations of the institution will be large enough that new business models will emerge. A company specializing in higher ed AI will be able to collaborate with schools to offer AI as a service. They will not charge up-front fees but instead will take a revenue share from increased enrollment yield and retention.
Campus development offices will partner with higher ed AI providers to increase alumni giving. AI could help in the recruitment and retention of faculty. In figuring out which new degree and nondegree programs to offer, AI will play a role in supporting a vast range of decisions of campus leaders.
The technology will not replace judgment, expertise and the ability to communicate the reasons behind decisions. At schools that use AI well, the number of jobs should increase (and become better jobs), as the business of the institution will grow.
The only way that higher ed people should fear AI is if other schools -- competitors -- realize earlier than those at your institution how artificial intelligence is set to change how universities run. AI will be a competitive advantage. A differentiator.
What companies do you think have a head start in becoming higher ed AI specialists?
Will the higher ed AI industry become specialized, with companies focusing exclusively on the challenges of higher education? Or will general technology providers succeed in building higher ed verticals within their larger AI businesses?
Will the AI revolution hit the learning side of higher ed in the same way that it will impact more business/administrative-related operations?
Are you with me in the belief that the evolution of narrow artificial intelligence (including big data and neural networks) will be as big a deal for higher ed as online learning is now?
