OPINION: When Stanford University threw open its artificial intelligence course to anyone with a web connection last year, it shifted the rock that triggered the avalanche. More than 160,000 people signed up from all over the world.

Earlier this year, the Massachusetts Institute of Technology and Harvard University joined forces to offer courses online, targeting hundreds of millions of students.

On many traditional university campuses it might seem like business as usual. However, higher education is in chaotic transition. What we do now to manage that change will determine the future.

We have an idea. Let's stop lecturing, let's outsource training to computer screens and let's concentrate on mentoring; that is, putting the really important face-to-face time with teachers to much more effective use.

Imagine if we could harness the production values and communication power of a Hollywood blockbuster such as Avatar, for example - or even the quaint appeal of Angry Birds - to inform and to train students more effectively, rather than to momentarily entertain. Or imagine a drop-down menu that would allow students to select a course, a voice of their choice, a digital avatar to talk to them and a pace and time frame to suit.

These aren't gimmicks to entice a lazy younger generation with short attention spans; they are effective, efficient and potentially interactive communication tools. The US-based Khan Academy, an online non-profit college, has already demonstrated that combining modest digital production values with sound pedagogy can make a real impact.

The academy has delivered 168 million lessons online.

If we consider the status quo, what is at stake is an outdated education system that we already know is not working well enough. In the US, for example, of the 600,000 college students who take first-year calculus every year, 250,000 fail. Studies of mechanical engineering outcomes paint a similarly depressing picture.

Tampering with the existing education system is not enough. We need to reinvent higher education. A good place to start is to distinguish between training and education. Much of what we currently call education is training, poorly done.

Digital media offers us an unprecedented opportunity to outsource training to computer screens. All courses are different, but to clarify: learning a CAD program is training, learning to design requires education, learning spelling and grammar is training, learning to communicate requires education, learning calculus is training, and learning to think using calculus requires education.

Every other industry is already harnessing the power of rich, visual digital communication. Hollywood spent about $500 million making Avatar - coincidentally, roughly the same amount we waste every year when 250,000 students fail calculus in the US alone. Why not collectively invest in, say, creating beautiful, photorealistic, animated content for biology that takes students right inside cells?

At the University of NSW, medical, science and engineering students are already using an adaptive e-learning platform invented at the campus to develop and hone the skills they need in virtual laboratories and clinics and by working with virtual patients and microscopes. These kinds of ed-tech tools are rapidly changing what we do on campus and online.

But, if we harness technology to inform and train, then how do we teach? In much the same way good teachers always have: by fostering apprentice-like interactions between students and experts. If we work alongside students as mentors to teach ''by doing'', we can achieve better outcomes than if we sit them down in a big hall, or even in a smaller classroom, and ask them to copy down abstract formulas.

Practical, hands-on exercises and competitions in engineering education, for example, were developed some decades ago at MIT and have since been replicated all over the world.

We can work widely, too, creating new products, challenging students with ''a job too big, a time too short, a team too large and a budget too small''.

They not only succeed but learn lots of relevant, real-world people and communication skills in the process.

Higher education worldwide has the teaching talent, technology and pedagogical insight we need to take on this challenge. But we need to avoid getting stuck in one trap.

This is not a debate that pits computers against teachers or online courses against the on-campus experience. We need the best of both worlds, education that is virtually and personally interactive.

No one yet knows quite what that will look like. But we can all get on with dispensing with paper textbooks, 50-minute monologues and passive learning as quickly and efficiently as possible.

Woodie Flowers is Emeritus Professor of Engineering at MIT and was a keynote speaker at last week's Adaptive eLearning Forum at UNSW. Dr Dror Ben-Naim invented the Smart Sparrow adaptive e-learning platform while leading a research team at UNSW.

This opinion piece first appeared in The Sydney Morning Herald