22 de janeiro de 2012

Will We Need Teachers Or Algorithms?


VINOD KHOSLA

Sunday, January 15th, 2012

robot teacher
Editor’s note: This is Part III of a guest post written by legendary Silicon Valley investor Vinod Khosla, the founder of Khosla Ventures. In Part I, he laid the groundwork by describing how artificial intelligence is a combination of human and computer capabilities In Part II, he discussed how software and mobile technologies can augment and even replace doctors. Now, in Part III, he talks about how technology will sweep through education.
In my last post, I argued that software will take over many of the tasks doctors do today. And what of education? We find a very similar story of what the popular – and incredibly funny! – TED speaker Sir Ken Robinson calls “a crisis of human resources” (Click here for the RSA talk from the same speaker which has been animated in a highly educational fashion). At the TED 2010 conference, he stated that “we make poor use of our talents.” Indeed, in the same way that we misuse the talents and training of doctors, I believe we misuse the talents and training of teachers.
I want to comment on what I consider a far greater misuse of talent and training: that of our children/students, mostly here talking about high school education. We have focused so much of our education system on children attending primary school, then middle school, then high school, all with the objective of attending university. This is a progression that still remains unchanged and largely unchallenged. Yet, this system is completely linear and, most tragically, unwaveringly standardized not only through instruction methods, but also through testing. Worse, it is mostly what I call “fixed time, variable learning” (the four-year high school) instead of “fixed learning, variable time” to account for individual students’ capabilities and status.
Identifying Emerging Trends In Education
There are new key trends that I see emerging in education enabled by advancing technology: namely decentralization and gamification. By understanding these trends, it is much easier to imagine why we won’t need teachers or why we can free up today’s teachers to be mentors and coaches. Software can free teachers to have more human relationships by giving them the time to be guidance counselors and friends to young kids instead of being lecturers who talk at them.  This last possibility is very important—in addition to learning, schools enable critical social development for children through teacher student relationships and interacting with other children—classrooms of peers and teachers provide much more than math lessons.  And by freeing up teachers’ time, technology can lead to increased social development rather than less as many assume.
Still, nearly all the attempts at technology in education have mostly failed so far, but I doubt they will continue to fail. I believe the failures have been failures of tactics rather than failures of strategy. In other words, just because some social networking sites like Friendster and Myspace failed does not mean that all social networking sites (like Facebook) will fail!
Let’s start with decentralization, which involves not only making content available online but also producing content that is interactive and mobile. It’s encouraging to me that we are starting to seriously experiment with content that is different than linear translations of books to online. With the new platforms, we have the ability to rapidly run experiments with new styles, techniques and resources (like social learning) which will lead to a new understanding of education.
At a very simple level, organizations like Khan Academy are making up for students who have bad teachers by starting with good lectures on every topic. And it seems to be working; hundreds of thousands of students are already accessing these videos, making up for what is lacking (likely from their “average teacher” – on the other hand good teachers, the top 20%, like great doctors, will always be in demand, and though each of us can tell stories about an awesome teacher, anecdotal counterexamples to my assertions are not “statistical proof” of the general quality of teachers). Meanwhile organizations like my wife’s CK12.org are making the basic content for high school education free and continuously improving (because they are open source).
This decentralization does not have to benefit only the students, though. Coming next from CK12, early in 2012, for example, will be lesson plans for teachers to access, “bionic software” to help them assess their students (Khan is also adding this to their system). The CK12 system in 2012 will also allow for students to teach themselves from a concept map of all 5000 or so concepts that each student needs to know to qualify to get into MIT or Stanford (the number of concepts depends on granularity and is mostly semantics). CK12 will allow teachers to guide individuals or for students to guide themselves while being aligned with state (or country) curriculum requirements using any of several different learning modalities (text, video explanations, experimentation, labs, playing with Flexmath, simulations, Q&A bank for students or teachers to self-test, social Facebook-like help for students, and teachers and much more). Systems such as these will enable near universal learning (again with some exceptions). There are numerous other examples.
The important thing is not the specific first instantiations of these systems but that they offer a customizable playground for low-cost experimentation.  Today, to try a new experiment in education in the US means starting a new school, training new teachers and taking years and tens of millions of dollars. My hope is that environments like that of CK12.org (and I see many more bubbling up, too numerous to mention here) will decrease the cost of experimentation and hence dramatically increase the amount of experimentation. This will result in accelerated innovation in education well beyond anything that has happened in the past.
The universal availability of inexpensive web access devices like tablets and smartphones and new trends like gamification and social software will surely add to the acceleration. Meanwhile the ability to collect data online, at a scale not possible in traditional systems will help us better understand student behavior. When every click and every hesitation at every stage of every reading, assignment or problem is available to analyze with big data techniques, we will finally understand at an exponentially faster rate what works and what does not.
The other trend I am excited about is gamification. When parents think of games, they usually consider them a waste of time. More importantly, they consider them a waste of talent. The debate about the place of games in learning rages on, but one aspect of gaming is unequivocally clear: it’s sticking around. Therefore, I firmly believe that we should embrace it and harness its best parts to drive the education of our children who grow up with online and mobile games. And I really mean, grow up with it! In a NYT article also from last year, according to research by the Joan Ganz Cooney Center, 60 percent of the top-selling iPhone apps on the education store are made for toddlers and preschoolers. Do we expect these children to relinquish and forget their app- and game-centered development after they get to first grade? This is completely unreasonable! And for me it is easy to envision how we can make education more engaging with these approaches, hence enhancing learning at all levels be it kindergarten or medical school. There is sufficient early evidence to suggest these assertions will be proven correct. I am a fan of views like those of Jane McGonigal whose  #1 goal in life (quoting her website) is to see a game designer nominated for a Nobel Peace Prize because of the impact game design can have on humanity.
I am particularly excited about the possibilities when high school education moves from teachers talking uniformly to bored A students and clueless D students, fifty to a class, to individual, gamified, and adaptive-difficulty systems, that leverage our social inclinations as demonstrated by Facebook. Imagine friends helping us understanding subjects while they also understand our context.  Both the students helping and the ones being helped are likely to understand the subject matter better in myview.  And with points and stars and badges and the like both are likely to want to spend more time participating, and will be more motivated when they do participate compared to today’s average classroom. Add reputation systems to that and one has the beginnings of a revolution. The content to train the trainers will be produced by some of the top 20% of teachers, and over time technology will multiply the impact and reach of these top teachers, motivating the rest of the best to participate as well. Other motivated teachers can feel free to jump in while the rest can go enjoy their favorite TV show.
Envisioning Future Education
Can you imagine an educational platform like CK12 version 9.0 in ten years, for example, with the excitement-generating, attention-grabbing, and skill-building potential of a Zynga game times ten? Can you imagine an environment based around a game? The awkward early prototype example of this (that will get much better in its multi-generational evolution) is Quest to Learn. Rather than pushing education on its students, the teachers pull the students into education through a game-like progression exploring 21st-century skills such as code-based problem solving, social media generation and integration, and design through games. The beginnings of  these future trends of educational institutions and platforms are, therefore, already in place.
One other critical piece in my vision of the future is (re)-discovering the potential of each student as just that – a student. Pioneering social experiments such as Hole in the Wall have shown us once more in explicit terms the incredible ability of children to learn if self-motivated. Children who have never seen – much less operated! – a computer, were able to learn how to browse the web, play games, learn the basics of a foreign language, and read manuals to the software in the computer. All of this within the timeframe of less than three months. Most of these tasks, they accomplished within hours of playing around with the machine!
More importantly, they were then able to teach themselves and others in their community. Children have the natural ability to learn and teach. With socialization, big data analytics and gamification as helpful tools, the future of education lies in providing children with an environment in which they can learn in their own way, at their own pace, and their preferred style/methodology/modality. I suspect they will still be able to meet any state or university curriculum standards. I could even imagine each university defining its own standards, providing the ultimate customization that no typical school today could. We may not need as many doctors as we have today but I suspect there is still a major role for the 80% of teachers who are not in the top 20%. They can provide the “human touch” and be mentors and coaches. Maybe teaching will become interesting enough to attract more teachers!
So is it possible to imagine solving the healthcare and education problems without doctors and teachers in their traditional roles within a decade or two or three? (Timing is always far off and the technologists always over-estimate the near term while underestimating the long-term because of the exponential nature of progress that builds on each previous step). As I’ve mentioned before, if computers can drive cars and master all the knowledge required to win Jeopardy, then surely it won’t be long before they will be able to diagnose disease and teach high school. With more and more data, these teaching and healing algorithms will keep improving and will free up human teachers and doctors to do what they do best.
Technology can allow us to make better use of our natural human resources, be they related to our health or to our education. Empowering patients to understand themselves better through continuous and comprehensive data and enabling students to develop themselves through accessible and attractive environments…this is the future I see. And if I can see it from these emerging trends, the key takeaway, then is, this: if we can see it, we can most certainly grasp it. All we need to do to reach this future is to invent it.

Nenhum comentário:

Postar um comentário