In July 2011, Sebastian Thrun,1 who among other things is a professor at Stanford, published a short video on YouTube, announcing that he and a colleague of his, Peter Norvig, would make their course “Introduction to Artificial Intelligence” available free of charge via the internet. By the start of the course in October, 160,000 people from 190 countries had registered. At the same time, Andrew Ng, also a professor at Stanford, offered his own course on machine learning free of charge via the internet, which was registered by 100,000 people. Both courses ran for ten weeks. Thrun’s course was completed by 23,000 people and Ng’s by 13,000.
Such online courses, featuring short video lectures, discussion groups for students, and systems for automatically grading their coursework, became known as Massive Open Online Courses (MOOCs). In 2012, Mr. Thrun founded an online education startup named Udacity, and Mr. Ng co-founded another, called Coursera. The same year, Harvard University and the Massachusetts Institute of Technology (MIT) jointly established edX, a non-profit MOOC provider, under the leadership of Anant Agarwal, head of MIT’s artificial intelligence laboratory. Some thought MOOCs would replace traditional university education. The initial hype surrounding MOOCs has since somewhat diminished (although millions of students have taken some form of online course). However, the MOOC boom highlighted the enormous potential for delivering education over the internet in small, manageable segments.
The fact that Udacity, Coursera and edX emerged from Artificial Intelligence laboratories highlights the belief within the AI community that educational systems need reform. Mr. Thrun says he founded Udacity as an “antidote to the ongoing AI revolution” that will require workers to acquire new skills throughout their careers. Similarly, Mr. Ng believes that given the strong impact of their work on the job market, AI researchers “have a moral obligation to stand up and address the problem they create”; Coursera, he says, is his contribution. Moreover, AI technology has great potential in education. “Adaptive learning” – software that individually composes lessons for each student, presenting concepts in the order they will find easiest to understand and giving them the ability to work at their own pace – seemed to be waiting in the corner for years. But new machine learning techniques may finally help realize its promise.
Economist: Re-educating Rita. Available here.
The above text is a part from an Economist article entitled “Retraining Rita”. It is touching that all the artificial intelligence labs of the most prestigious universities in the US took the same humanitarian turn shortly after 2010, as described in the article. Perhaps then came the moment of realization and they said “Yes, we must admit that artificial intelligence will cause Joe the worker to lose his job! We created this problem. Therefore, we will be the ones to fix it. And we will take immediate action for this, so that from now on we can sleep peacefully.” Thus, during those years, they created the major providers of massive open online courses and since then they have been helping humanity and making the world a better place with more education for all. Absolutely!
MOOCs are a point in the history of an effort to shape a form of massive (and at the same time personalized) electronic education. The pioneers of massive open online courses, at the beginning of their creation, proclaimed that they aimed to educate the masses without discrimination, regardless of origin and economic capabilities. MOOCs were presented as a way to provide access to high-level education to those who could not have it. When this type first appeared, it attracted intense interest for several reasons. On one hand, because access was “open” via the internet, so one could attend classes for free, and moreover from the most famous universities in the world. And on the other hand, because they were perceived as the rival fear of the traditional education model for the 21st century.
It is not clear to us why, a decade ago, this explosion of MOOC creation and platforms hosting them occurred. What is certain, however, is that significant investments were made in this field by governmental and non-profit organizations, institutions such as the Bill & Melinda Gates Foundation, and other investment funds.2 Perhaps one reason is that major academic institutions are seeking ways to follow the trends of the digital age. The internet is filled with unregulated knowledge and instructions on virtually everything, allowing us to learn whatever we need without academic involvement. Therefore, shouldn’t the organizations that traditionally serve as the “bearers of knowledge” be able to exercise some control over how this knowledge is certified?
However, the noise created around MOOCs gradually faded, especially as time passed and it became clear that their title does not actually express what they ultimately are. It was discovered that the completion rates of enrolled students are very low, under 10%, while the overwhelming majority of MOOC students are already degree holders, employed, and come from wealthy countries. Later, most MOOC providers, including Coursera, edX and Udacity, imposed fees, if not necessarily for access to the courses themselves, at least for obtaining a certificate of completion after successfully finishing the course. Thus, one could say that these are simply a form of paid online seminars. But this does not seem to have bothered anyone. Neither the creators who had “so many dreams” for free online education, nor those who take the online courses. And obviously, no one protested about the loss of free massive online education…
MOOCs therefore proved to be neither massive enough, nor open enough. But they certainly were online courses. The MOOCs title was a good advertisement. But the issue was not simply the creation of yet another successful educational product. Rather, it was more of an opening of education towards the new possibilities offered by the world of data for the utilization and optimization of “human capital.” Also, it functioned as an experiment on a large base of learners regarding the impact of new learning methods for specific subjects (and not entire study cycles) and the use of digital tools for assessing their performance.
data collection and utilization
Entrepreneurs in education (also known as edupreneurs) refer to various benefits that online courses can offer to businesses and for enhancing technological development. One of these is that online course platforms provide the ability to collect large databases (big data) regarding the ways people learn. This is a massive amount of data that would not be possible to gather within the framework of traditional education. The data can be analyzed and studied in order to develop educational software techniques and make them more efficient and personalized for each learner. To such an extent that they can replace a personal teacher in terms of presenting educational content, as well as evaluating performance.
Indeed, precisely these personalized learning methods are considered the next major milestone in e-learning, encompassing both higher education or professional training and basic education, that is, schools. Adaptive personalized learning refers to the use of technology (computers or other devices) with the purpose of adapting the lesson to the needs of each learner. Such methods are promoted as more effective for enhancing performance, and in recent years, significant investments have been made in this direction.3 The introduction of appropriate technology and algorithms is considered necessary in order to apply personalized learning on a large scale. Of course, the lesson will “adapt” based on the data with which each learner feeds it through her own actions, using suitable machine learning algorithms. Thus, regarding the development of this particular field, data analysis and artificial intelligence play a central role. Interactions with the software and the user’s knowledge acquisition “patterns” from the electronic platform are used as data that train the machine learning algorithm on the individual “peculiarities” of the user by creating statistical models.
MOOCs have made a significant start regarding the use of data analysis technologies in education. As stated in an article from MIT Technology Review titled “The Crisis in Higher Education”:
What specifically excites Thrun [note: the creator of Udacity] and other computer science experts about free online courses is that, thanks to their unprecedented scale, they can generate massive amounts of data required for effective machine learning.
Daphne Koller [note: a professor of computer science at Stanford and one of the creators of Coursera] states that Coursera has developed its system through intensive data collection. Every change in a course is tracked. When a student pauses a video or increases its playback speed, this choice is captured in Coursera’s database. The same applies when a student answers a quiz question, revises an assignment, or writes a comment on the forum. Every action, no matter how insignificant it may seem, is fed into the statistical mill.
The collection of information about student behavior at such a microscopic level of detail, says Koller, “opens new avenues for understanding learning.” Hidden patterns in how students navigate and ultimately master complex topics can now come to light. [..]MIT and Harvard design edX to be both a tool for education research and an electronic educational platform, says Anant Agarwal [note: the CEO of edX]. Researchers are already beginning to use data from the system to test their hypotheses about how people learn, and as the course portfolio grows, so will the opportunities for research.
https://www.technologyreview.com/s/429376/the-crisis-in-higher-education/
Personalized learning, which artificial intelligence can facilitate through the analysis of behavioral data, preferences, and performance of each learner, can indeed be more effective in terms of mastering what one is required to learn. However, does adaptability refer to the algorithm in relation to the student, or is it the other way around? The algorithm will guide the student toward what she needs to learn. Therefore, ultimately, it is the student who must bring her own thinking—through her personal approach—to the point that the algorithm attempts to lead her to. Thus, personalized e-learning can offer as much freedom as a video game does: its world may seem open and allow you to wander around, but in order to progress in the game, you must complete specific quests or missions. When a machine intervenes in the process of the game or learning, the rules set are strict and well-defined, even if the algorithms it uses appear to be fed by “individual peculiarities” and adapt to them. This is a relationship in which the algorithm holds a position of power—even though it gives the impression that it doesn’t even exist, simply running in the background. Ultimately, a personalized learning system aims to make the student measurable and controllable by the machine, with the purpose of increasing the effectiveness of her education.
In the context of electronic personalized education, each learner will have a digital skills profile that will be composed of multiple variables and data, and will determine how they will proceed at each subsequent stage.
wouldn’t it be better to leave education as it is?
Facing a new form of education, in electronic format, we do not defend the old one. Moreover, at this stage, the new educational methods are evolving from those that already exist and are built upon their own foundations. From one perspective, modernization through the introduction of online courses is logically supported by smaller or larger universities around the world. Even if the technologies behind online courses point toward a radical transformation of the teaching profession itself.4
As the use of new technologies becomes more widespread and intense, the old mass educational system, as it was shaped in the 20th century, becomes increasingly obsolete. Its methods are now outdated and the content of teaching appears more and more obsolete in a world where forms of work are constantly changing through the introduction of machines and automation. Even “intellectual” work, which was believed to be carried out exclusively by the human mind, is now being mechanized to an ever greater extent. Many “intellectual processes” (simple or complex) can now be automated quite easily and completed with greater speed and accuracy using the mechanical “thinking” of algorithms. Thus, the cultivation of broader “intellectual abilities,” which is the responsibility of the educational system, is losing its importance. Meanwhile, the possession of very specific skills or characteristics and specialization in particular digital tools is gaining greater significance.
Academic institutions, however, are not abandoned. Because they are endowed with a magical promise: the papers, the certifications, the diplomas! Imagine what would happen if suddenly the educational system gave no diploma, not even a certificate of attendance; then it would be considered secondary, if not meaningless. The luster that still sustains it would fade. Perhaps it would be treated like public libraries are treated now. Schools and universities would simply exist and address those who want to study. Most people wouldn’t even know where the neighborhood school is located. In such a case, therefore, the educational system could only be maintained as compulsory. But how could compulsory education be perceived when it provides no credentials to learners that they can use for their personal advancement? Rather, as a purely disciplinary mechanism…
The educational system, of course, has not stopped issuing diplomas. However, as time passes, these pieces of paper are increasingly no longer considered something great. Some forgotten remnants of the previous century complain that diplomas are being downgraded and demand “diplomas with value.” They are outraged by the generation of unemployed graduates, of scientists who do not work in their field and do not receive compensation commensurate with their education. It’s a shame for all those who believed and continue to believe in the ideal of the professional graduate, an ideal that has been crumbling for two decades now.
And of course, the devaluation of degrees is not something that happens only in our parts. An article by cnn titled “Do we need a revolution in higher education?” describes obtaining a degree as a “bad investment”:
In some cases, a college degree can no longer guarantee potentially high lifetime earnings, as it once did. An online salary ranking system named PayScale.com calculates the 30-year return on investment of a student’s education at the 1,300 top colleges nationwide based on graduates’ average salaries and tuition costs. A 2012 report of theirs shows that out of 4,500 colleges and universities in the U.S., only the top 800 to 850 provide a return on investment higher than 4%. In purely economic terms, students would be better off investing their tuition money in stocks rather than spending four years at many of our national colleges.
http://edition.cnn.com/2012/06/13/opinion/bennett-higher-education/index.html
education of “human capital” in the digital age
Experts agree that the old educational system does not meet the needs of modern capitalism. In the digital world, they see opportunities to direct the education of “human capital” in order to cover capital’s demands regarding the qualifications it wants future workers to have. And these “qualifications” on which people should be educated can range from specific knowledge and handling of certain tools to elements of character. A key advantage that digital education is considered to offer is the creation of a modern form of lifelong learning that will better enable employers to select their employees’ profiles based on the individual skills they possess. This lifelong education will be able to take place during free time and will not need to follow a specific program, as is the case, for example, with seminars and postgraduate programs. At the same time, the responsibility of each individual to maintain their personal capital during free time is becoming increasingly urgent.
Continuing the translation of another part of the article “Re-educating Rita” with which we began this text, we encounter the following positions of the “experts” regarding the overcoming of the old education system and the new opportunities they foresee for transforming the educational system through its digitization, so that it will be more efficient for modern capitalism:
Adapt and survive
[..] Even outside the Artificial Intelligence community, there is broad consensus that technological progress—and artificial intelligence specifically—will require major changes in the way education is delivered, just as the Industrial Revolution did in the 19th century. As factory jobs replaced agricultural ones, literacy and numeracy became far more important. Employers realized that more educated workers were more productive, but were reluctant to train them themselves, since they might be poached by another employer. This pushed for the introduction of general state education based on a factory model, with schools supplying workers with the appropriate skills for factory jobs. Industrialization thus transformed the need for education and simultaneously provided a model for delivering it. The emergence of artificial intelligence could do the same thing again, making the transformation of educational practices necessary and, with adaptive learning, providing a model through which this can be achieved.
“The old system will need to be revised very seriously,” says Joel Mokyr of Northwestern University. Since 1945, he notes, educational systems have encouraged specialization, so students learn more and more about less and less. But as knowledge becomes obsolete more quickly, the most important thing will be to learn how to re-learn, instead of learning how to do one thing very well. […] In the future, as more tasks become automatable, the tasks where human abilities will be of greater value will be constantly changing. “You have to keep learning for your whole life – this has been evident for a long time,” says Mr. Ng. “Whatever you learn in university is not enough to carry you forward for the next 40 years.”
Education should therefore be interwoven with a full working schedule. “People need to continuously learn new skills to stay current,” says Mr. Thrun. Therefore, his company focuses on “nanodegrees”5 which can be completed in a few months, alongside a job. Studies for a nanodegree, for example, in data science or web programming, cost $200 per month, but students who complete the course within 12 months receive a 50% refund. […] Some companies, such as Udacity, charge per course; others, such as Lynda.com, which is owned by LinkedIn, a business networking site, charge a monthly fee for access to all courses. (It’s not hard to imagine LinkedIn comparing users’ skill sets with those required to apply for a specific job—and then offering users courses needed to fill their gaps.) Users and potential employers sometimes find it difficult to determine which skills have real value. More collaboration between government, educational providers, and employers on certifications would help.
[…] As job-related skills come and go, having solid foundations in basic reading and arithmetic skills will become even more vital. But teaching “soft” skills is also becoming increasingly important. In a 2013 paper, James Heckman and Tim Kautz of the National Bureau of Economic Research in America argue that more emphasis is needed on “character skills” such as perseverance, sociability, and curiosity, which are highly valued by employers and closely related to workers’ ability to adapt to new situations and acquire new skills. Character is a skill, not a trait, they say. […]
This effort, which does not originate solely from specific universities or companies, but is centrally promoted by the states of the capitalist developed world. A basic objective is the creation of a more efficient educational model that will shape the capabilities of the workforce in such a way as to serve the needs of capital. Also, the shift of state responsibility for education to individual obligation for maintaining and developing one’s own knowledge capital is already taking place. Each individual, as an entrepreneur of themselves, should take care to follow developments, and it should be ingrained that failure is an individual responsibility. Furthermore, there is an evident attempt to organize the untamed knowledge available on the internet, in order to make it functional (i.e. profitable) both for the educational market and the labor market. The ability to present skills without proving them through some academic title should be considered a given. However, this could not remain so for long, because realizing such a possibility would lead to questioning the necessity of certifications, as well as the distinction between “skilled” and “unskilled.”
Under these conditions, it is attempted to establish as a personal obligation for each individual to self-educate and accumulate degrees, even while employed, so that their labor, as a commodity, maintains its “competitiveness.”
how much more investment in human capital?
The game over gathering from the cinematic side makes a systematic critique of the educational system and talks about the decline of the traditional model since 2010. And with this realization: that the educational system and the degrees it distributes are outdated, we could proceed to question the illusions that accompany the collection of degrees.
And yet, despite the evident obsolescence of old certainties, faith in the value of degrees is devoutly maintained. If one degree is not enough, we will take a second and a third… Master’s, doctoral, etc. And we will hope that for us the future will have better prospects and we will continue to struggle to survive amidst the inflation of certifications. We have seen the effects of this ideology: 30-year-old “kids” living in their parents’ house to study and collapsing psychologies of graduates who have “invested a lot” but cannot find jobs in their field. A kind of mass social melancholy is the result of tying hopes of social ascent to a dead educational system.
The experts are therefore coming to propose a digital transformation, which will be capable of preserving the same manic-depression in electronic packaging. Are you afraid? Are you worried that you won’t find a job or will be laid off from your job? You can now collect lifelong nano-degrees, micro-master’s degrees, even in parallel with your job! And not only should you not resent it, but you should actually like it. What better thing do you have to do in your free time, than to build your digital profile and invest in your personal capital?
As long as we do not question the logic that wants us to be small investors of our individual capital, we will continue to fall into the trap of the proliferation of certifications, old or new. Technocrats like Thrun know that the ability to acquire knowledge has escaped the boundaries of university campuses. That is why it is crucial to initially regain control of the validity of knowledge provision from its traditional carriers. Moreover, the lifelong pursuit of a continuous upgrade of skills, especially for those working in environments where production depends on and is controlled by information machines, seems to partly respond to the “highly sought-after” alignment of education with the labor market. However, such a pursuit, although they try to convince us that it is a sure path to “advancement,” ultimately remains nothing more than an update of the old careerism of the 90s: many will try it, few will succeed. With one significant difference, however: increasingly more people are trying, and increasingly fewer are succeeding.
Could a capable portion of the “lifelong learners” around the world confront and utilize modern knowledge acquisition capabilities outside the interests of businesses and markets, outside the narrow conception of individual capital?
Shelley Dee
cyborg #08 – 02/2016
- Thrun has been a public figure since 2005, when he converted a Volkswagen Touareg and won a competition for driverless cars. The sponsor of the competition was the US Department of Defense. In 2007, he led Google’s program for developing self-driving cars. He also founded Google X, the laboratory where Google Glass and other research projects were developed. ↩︎
- Graph presenting investments and donations to MOOC providers from the newspaper “The Chronicle of Higher Education”. Available here. ↩︎
- An example of this is the couple Mark Zuckerberg (CEO of Facebook) and Priscilla Chan, who announced at the end of 2015 that they would donate 99% of their Facebook shares—worth 45 billion dollars—to a series of causes, with the primary focus on developing software “that understands how you learn better, and where you need to focus.” As they themselves say: “The initial areas we will focus on will be personalized learning, curing disease, connecting people, and building strong communities.” Let’s assume that the last two mean more Facebook for everyone? ↩︎
- You can read about the topic in the article: “The (partial) digitization of the analog educational system”, in cyborg #7. ↩︎
- Accordingly, edX offers “micro-masters”, that is, micro-master’s degrees. ↩︎