The concept of a learning society has been around since 1973, when Donald Schön wrote that in order for societies to adapt and evolve with the increasing rate of change in modernity, learning must be conceived as a lifelong activity for individuals and groups, not restricted to institutionalized formal learning. The learning society rejects the fallacy that learning ends with childhood, often advocating for varied technological systems that enable informal, self-directed learning opportunities for learners of all ages. By positioning this existing conception in a future age of non-human learners, this manifesto aims to augment the concept of a learning society, calling for a rise in global digital salon culture.
At present, affluent society is witnessing and participating in the education of non-human “intelligence” in the form of trainable algorithms built into the products, systems, and experiences of daily life. “Machines”, more aptly called simply algorithms, feed on almost limitless masses of data, using statistical learning techniques, such as “decision trees”, to recognize patterns in data and apply these learned trends to new sets of data. In result, the ecosystem of learners, teachers, and educators (both formal and informal) in society is expanding to encompass non-human learners and other species of intelligence which we interact with on a daily basis.
Since the appearance of Maria, a conscious technological being, in Metropolis in 1927, a mythology of consciousness and intelligence in technology has materialized in various forms of media. The most recent examples of this in modern media include Samantha of Spike Jones’ Her and Ava of Ex Machina. These representations of “intelligent”, learning technology transpire in the form of female human-like conscious beings.
However, machine learning technology lives invisibly within everyday digital product experiences like Amazon product suggestions and Facebook ads curated based on your Facebook activity and interests. The products we use on a daily basis are constructed of statistical learning algorithms, which is the closest that society gets to “artificial intelligence” at present.
During an internship last summer, I attended a one day conference— called the Inferences Summit— that was entirely devoted to the idea of “inferences” that systems like Cortana (Microsoft’s “Siri”) gather from users’ locations, social interactions, preferences, etc.— information collected from our smartphones and other technological prosthetics. From this extensive data, the algorithms embedded in our products “learn” about our habits and patterns in our lives. Ultimately, I walked away from this conference slightly creeped out about how many “inferences” machine learning technology (silently) attempts to make about us everyday.
However, if one was to ever begin worrying or feeling threatened by the capacities of machine learning technology today, apps like AI Scry offer a form of solace through an entertaining interaction with AI. The app attempts to recognize objects in reality, based off of mass amounts of data that it was “trained” or “educated” with. Frequently, such as with the rock and sea horse, the AI is vastly incorrect— often, it thinks rocks are bananas or seahorses are humans snowboarding.
So as we can see, there still remains a colossal gap between the capabilities and characteristics of technology embedded in our everyday experiences today and the sensationalized portrait of learning machines in modern media. Within this gap lies great discourse around what it means to be a human or non-human learner. The emergence of technological “intelligence” calls for a reevaluation of our role as learners and educators of both human and non-human learners. One of the main divisive differences between machine and human learners is that learning for humans is intrinsically linked to giving meaning and purposeful to livelihood. Learning is often a means to an end, and that end– purpose, principles, and intent– and humans determine the purpose, principles, and intent of learning for both human and non-human learners.
However, there’s still a large group of highly regarded technologists who have incredibly pessimistic and fatalistic perspectives about the relationship between human and non-human learners in today’s society— Bill Gates, Stephen Hawking, and Elon Musk are just a few. This presentation will follow this same structure— I’ll begin with a heavily pessimistic vision of what learning societies could look like in an age of non-human learners, and I’ll finish with a more optimistic illustration.
Let’s imagine that in five decades from now, the United States government has decided to end all forms of privatized education (private K-12 and higher education institutions), pulling inspiration from other democratic socialist countries in the world, like Finland. In Finland, private schools are illegal, which results in significant socioeconomic diversity in students in schools, since all children must attend public schools. With affluent families attending public schools, Finnish public schools receive the funding, support, and concern it deserves, so that all children obtain equally exceptional education from an early age. Now, the United States wants to attempt this approach to equality in education, realizing that allowing private stakeholders to treat learning like a business opportunity might not be the best ethical societal strategy.
In addition to the end of privatization of education, the rise of machine learning throughout society has caused a major shift in the way humans complete transactions. In this future society, all former forms of monetary currencies are obsolete, with simply personal data becoming the new currency. The State provides “free” education for all, yet a student must agree to give over all personal data collected throughout his or her learning experience as compensation to the national government.
Now, all tangible objects in society are connected through a technology that was developed around 2015 called the Internet of Things. Although those who lived during 2015 scoffed at the ridiculous concept of having all objects and environments in society become “smart”, the government realized this “ambient technology” approach to education seemed like a utopian alternative form of assessment to standardized testing. Extensive research in the early 21st century began to reflect the harmful effects of standardized testing on a child’s cognitive, social, and intellectual growth, so ambient technology gathered mass amounts of data on a child’s development as he or she interacted with objects, furniture, and learning materials in augmented learning environments. This silent assessment, in theory, would not disrupt a child’s learning process with high anxiety test-taking situations, allowing them to build critical problem-solving skills that standardized testing did not exercise.
A level-playing field of opportunity— where all children experience equally exceptional public education— combined with an end to standardized testing seemed like an utopian approach to education reform. The public initially responded positively to these major systemic reforms, until the negative consequences of a data-centric society began to surface. After a decade of ambient technology in classrooms, machine learning technology knew more about children than their parents, teachers, friends or they themselves did. Data was gathered on students from age 3, and entire behavioral patterns could be predicted for each individual child. Actions, reactions, and emotive states could be anticipated by algorithms, resulting in a highly predictable human population. With an increase in known and anticipated behavior in humans, society’s systems of power (governmental, economic, etc.) began to categorize people early in life, placing them on a spectrum of expected “success” achieved later in life. Students were “sorted” into separately curated courses, based on their expected chances of attending colleges, which was determined by data collected from the onset of formal education. Excess data on human learning and intellectual “progress” ultimately resulted in a destructive social system rooted in a deterministic future of anticipated behavior, heightened surveillance, and full reliance on technology for analytical skills.
This dystopian vision of a fully data-centric relationship with machine intelligence and learning can be avoided if we revisit the idea of a learning society, and expand upon this conception of lifelong, informal learning between social individuals. Initiatives for a learning society have greatly increased since the advent of the Internet— informal self-directed learning opportunities are bleeding into digital communities and these digital interactions are facilitating in-person learning opportunities as well.
Khan Academy and Duolingo are just two of innumerable examples of informal self-directed learning opportunities that exist solely online or in mobile applications, providing individual learners with autonomy to set a learning pace and curate specific learning topics. In the realm of socialized self-directed learning, General Assembly, Creative Mornings, and Meet-ups represent just a tiny fraction of in-person learning communities organized separately from formal education systems or workplace-initiated programs. A rise in communal maker spaces, shared shop facilities, and resource sharing services (such as 3D hubs) are empowering individuals to collectively learn and share skills outside of formal systems of learning. The MIT Media Lab Lifelong Kindergarten research group is developing a number of tech-enabled community learning experiences for adults of all backgrounds, encouraging creative behavior beyond the years of colored blocks and crayons in early education.
Because these existing examples of decentralized communities of learning reflect a more locally focused approach to a learning society, this manifesto calls for a more globally connected initiative for learning societies in the near future. In addition to the emergence of local learning societies, society is in need of transnational accountability in learning and cultivation of a desire to continue learning throughout one’s lifetime.
Prevalent in the 17th and 18th centuries in France, salons acted as social hubs of intellectual exchange and discourse among attendees, inviting both men and women to the literary and philosophical movements of that era. In today’s society, these informal gatherings are re-emerging in the virtual expanse of the Internet, in the form of participatory publishing platforms like the MIT Media Lab’s PubPub or Membrane, an experiment by the New York Times Research and Development Lab. In-person salon culture, still enabled and amplified by networks of users, is embodied by large conferences such as TED or the resurgence of community-organized book clubs, meet-ups, etc.
However, many of these digital and in-person initiatives are still focused locally, and in initiatives like TED, only an elite few voices are heard. That being said, translation technologies today are beginning to break down language barriers that might have inhibited disparate groups of people around the world from communicating in the past. Machine-learning-enabled technology are opening doors to global communication, allowing grassroots organization of learners throughout the world to communicate. Consumer software like Skype translator would allow decentralized initiatives between groups across cultures to emerge, cultivating a global digital salon culture to (re)emerge.
This newly manifested salon culture in the near future could only occur if individuals across cultures actively work to erase the “filter bubbles” in our personalized web experiences that heavily curate and personalize, yet restrict what we actually experience on the Internet. Algorithms behind these web experiences work to provide us with search results, suggested links, and suggested content that seem most relevant to us as individual users, based on past web behavior and our digital profiles. However, these highly curated experiences actually disconnect users from information about differing viewpoints, contrasting cultures, select world events, and so on. In a utopian global learning society built around a digital salon culture, bottom-up innovation by groups of individuals in various cultures, independent from business or government, would begin to dissolve (or rather, expand) these filter bubbles so that global communication between individuals could emerge.
This global learning society built around digital salons would break down these crafted virtual spaces of homogeneity with the help of NGOs and other decentralized organizations formed to aid in this effort. With increased connectivity throughout the world, exchange of ideas through online “salons” would begin to break down barriers between cultures, allowing people to learn directly from specialists in various disciplines. In addition to laymen and experts in disciplines communicating through salons, greater awareness of global political, social, and economic movements would be heightened with an increase in communication between publics. Content of global digital salons could expand beyond intellectual debate into topics of societal social change, turning salon culture into an additional mechanism for mobilizing grassroots activist efforts for change.
Presentation PDF final_manifesto
Ball, Christopher. “The Learning Society.” Royal Society for the Encouragement of Arts, Manufactures and Commerce Stable 40 (1992): 380-94. Web.
“Lifelong Kindergarten.” Lifelong Kindergarten | MIT Media Lab. MIT, 2016. Web. 16 May 2016.
McClellan, John. “Envisioning Learning Societies Across Multiple Dimensions John McClellan.” Global Stakes: The Future of High Technology in America (n.d.): n. pag. Web.
Miller, Vachel. “A Learning Society Retrospective.” (n.d.): n. pag. Web. 15 May 2016.
Pariser, Eli. “Beware Online “Filter Bubbles”” TED2011. TED, Mar. 2011. Web. 14 May 2016.