Reimagine Open: Building Better Internet Experiences
The internet is at a crossroads. After decades of rapid growth and global contribution to the human condition, today many are questioning the health of the internet. At such a time it is natural to examine the first principles that many consider key features of the internet as a medium.
The internet’s success and impact is most often attributed to a set of design choices and embedded values from the internet’s earliest days, commonly grouped together under the label of “the open internet.” Historically, this set of technical and architectural features distinguished the early internet (and later the World Wide Web) from other communications media like the telephone, broadcast television, or print. Some of these features were codified directly in technical standards. Others have served as design values or norms in the development of internet-related technologies, including the web.
Today the internet has moved far beyond narrow uses into fundamental infrastructure that is a central part of everyday life for at least half of the world’s population. In under thirty years, it has brought access to information, economic opportunity, and greater empowerment to billions of people.
Yet today’s internet is not the internet we want. A rising tide of harmful content, surveillance capitalism, security risks and other problems have shaken our collective sense of optimism that the web is truly an inherent force for good. A ubiquitous network, personalized data targeting, and new monitoring technologies – divorced from proper oversight – have the potential to enable a state of total surveillance at the hands of private actors and governments. Marginalized communities have been targeted online and experienced violence and intentional harm. Today people can do much more online, but not without putting their privacy, security, and individual autonomy at risk.
We often hear that openness or being “too open” is to blame for this situation. It is true that the builders of the early open internet assumed that the technical features of the network would “route around” problems of centralization and control, rendering legal or societal intervention ineffective. An early optimism that the system might evade many of the unwanted constraints of society has been replaced today with a realization that the internet is all too susceptible to societal and human frailties, with limited avenues for accountability or redress. “Open” has itself been co-opted by a wide variety of players to defend practices that once were viewed as antithetical to open values (including closed standards, walled gardens, or opposition to principles of network neutrality.)
Is openness an outdated concept? Are the internet’s failures today an unavoidable outcome of the open principles it was based on? Or is openness still a valuable concept to draw from as we chart a future path for the internet?
This paper traces the evolution of internet openness and the erosion of these values. And it examines how we might reclaim openness to address these real concerns about online harms while preserving the best features of the internet today.
Section I: The Early Internet and the Power of Open
Historically, a set of “open” technical and architectural features distinguished the Internet from other communications media like the telephone, broadcast television, or print. Some of these features were codified directly in technical standards. Others served as design values or norms, manifested in areas like the open source and free software movements. They included:
- Decentralized structure: A widely interconnected network of networks, which enabled connection without the need for third parties approval and radically improved access.
- Layered software model: A level-based model of abstraction, ensuring implementations at one layer do not impact other layers and enabling greater innovation and participation.
- “End to end” networks: Pushed intelligence towards the endpoints to minimize reliance on particular features of underlying networks, in turn enhancing innovation at the edge.
- Open standards: Publicly available, non-proprietary standards allowed developers the opportunity to create products and services offering consumers more competition and choice.
- Open source: Open source licensing offered greater access to tools, empowering people with fewer resources to problem-solve and create for themselves.
These early design choices translated into a set of consumer and developer experiences and possibilities that reinforced (and were reinforced by) a set of social values, including:
- Access: the capability to reach the network, information, audiences, markets, and each other.
- Opportunity: the chance to participate in the network as an individual with minimal restraint, without gatekeepers, and with the ability to resist centralized control.
- Empowerment: a broad-based ability to create, shape, and participate in online experiences – in contrast with the pre-existing model of broadcast TV, where the only choice was what to consume.
Openness never meant the absence of all restrictions. The open Internet developed against the backdrop of technical features, social norms, and laws that influenced human behavior on the network. Together these created a set of accountability mechanisms that provided a backstop and framework for a successful human experience of the Internet.
Healthy communities, with meaningful participation on equal terms, were core to the successes of the early internet. In that era, however, the community was small – a small set of technically savvy people actively engaged in building and operating a new medium. That community had a shared purpose, personal relationships, and networks of trust. Accountability was achieved as much through personal and direct connection as it was through laws and regulations (although those had started to emerge as well).
The early accountability mechanisms were also designed for a different scale of network. External attacks on the system could be thwarted by “routing around” the problem – e.g. if an IP address got blocked or a computer taken over, the network would adjust and was designed to be resilient to individual points of failure. Actors within the system were understood to be unlikely to harm the system because they were invested in the health and success of the network. Harm to one would be viewed as harm to all.
Section II The Open Internet Comes of Age
As the open internet scaled from millions to billions of users, and new business models and network effects changed its landscape, the fault lines in openness and accountability mechanisms began to show. Two broad diagnoses offer explanations of the crises that today afflict the internet: an erosion of technical and architectural features that uphold open values, and the inability for accountability mechanisms (like law, policy and regulation) to adapt and scale to meet these challenges.
The erosion of open: Today many of the original open systems have been deeply encapsulated and controlled by layers of proprietary computing systems. There is more open source software than ever before, and open source continues adding positive value to the overall ecosystem. And yet the experiences of many internet users have migrated to closed systems, most prominently today’s popular social platforms. These platforms do not share the same design principles as the internet infrastructure on which they were built – and don’t reinforce the same social values. Many of these platforms still use the language and rhetoric of openness to describe their activities and ethos, but these are often shallow commitments. Key aspects include:
- “Open APIs” but walled gardens? Today developers live in a world of “open APIs.” Many of these open APIs are a crippled substitute for an open standard. There may well be cases where APIs are justified instead of an open standard. However, as a default, open APIs offer a limited subset of the elements that make open standards a part of a healthy ecosystem.
- From open source to proprietary software: A similar change has occurred with open source. Today open source is mainstream; Google, Amazon and Facebook all use vast amounts of open source software to power their business. But each surrounds the open source elements with successive layers of proprietary and secret systems. As a result, these companies harvest the benefit of open source, but provide a limited set of broader benefits to other actors.
- “Community” co-opted: Over time, “community” has evolved from the open source ethos of a self-governing community aiming to produce technology that is usable by all on equal terms. Today it is commonly used to mean nothing more than “consumers who use our products.” When a single corporate decision maker determines what is possible for users, and how those “communities” are tracked and surveilled and sold to advertisers – it is not the empowering and participatory nature of the communities that built the web.
Accountability falls short: The existing accountability mechanisms for the giant closed platforms that dominate the landscape today are not fit for purpose. This problem has manifested for a few reasons:
- Unanticipated harms: The seriousness and scale of privacy harms, security risks, and abuse online (to name a few) have been difficult to anticipate, even for those building them. This inability to predict and adequately forecast the range of threats to users means that both regulation and technical solutions has been resigned to catch-up. As a result, even efforts to rectify the situation have been tilted in favour of maintaining (rather than disrupting) the status quo.
- Complex solutions: Even as the problem has become clearer, the question of possible solutions often prematurely ends because “it’s too complex”. This in part demonstrates the competing values involved in many internet policy debates today. But the resulting inertia has meant insufficient protections for internet users in the meanwhile.
- Corrosive business models and incentives: Whether privacy or misinformation, the root cause of many of today’s challenges is the business model that powers many of these platforms. Increasingly referred to as “surveillance capitalism”, the financial incentives of businesses has been to service advertisers. There has been serious and long-standing push back to regulation and any other accountability mechanism from private industry.
Beyond the private sector, accountability mechanisms must apply to government and intelligence agencies as well, who will inevitably be reluctant to move regulation against themselves.
Case Study
Reimagine Open In Practice – Illegal Content: How can we address the real concerns felt today about illegal content online, while preserving the best features of the internet as a foundation for communication and inovation? That is a question gripping governments around the world, and particularly the European Commission as it considers a proposed Digital Services Act that will shape online regulation worldwide.
The internet’s open architecture, coupled with the legal backdrop of the EU E-Commerce Directive and other innovation-friendly policies, has made it a natural engine promoting access to knowledge, greater civic discourse, and economic opportunity. Yet the internet’s very openness has been exploited by bad actors, and the persistence of illegal content online in particular remains an untenable threat to the internet’s original vision, jeopardizing both the legal and cultural environments that fostered broad engagement and growth. In that context, Mozilla has proposed a set of policy changes to create greater accountability for the largest platforms around concerns about illegal and harmful content:
- Greater transparency around advertising and algorithms that amplify content.
- Procedural accountability to incentivize strong moderation practices, using a risk-based approach that assigns greater accountability for those with the largest reach.
- Reorienting the business models that intensify harmful content dissemination, starting with privacy protections and new business models to attack surveillance capitalism.
Section III The Future of Open: 2020 and Beyond
Today, society confronts powerful and accelerating technological change that is simultaneously beneficial and dangerous. As we seek solutions, we can be guided by the core values of an open internet — a system for access, opportunity, and broad-based empowerment. It is these values that made the open internet such an attractive medium to people and such a powerful engine of innovation. And these values are still relevant.
But the original understanding of the open internet from its earliest days is not sufficient. Our new conceptions of openness need to reflect the lived experience of the internet to date, and with the diversity — and perversity — of humanity in mind. We should also reject the co-opted and limited model of “open” that has more recently taken root.
Achieving this better vision of open will require work on multiple fronts:
Systems design: Technologists and the technology industry need to build alternative products and experiences that return to openness principles. This is easier said than done. Current business models are wildly profitable, and closed, integrated platforms have been able to build ever larger domains. However, consumers have been sendings signals into the marketplace that open products can be successful as well. And many technologists long to build products without the harmful consequences of closed systems. But good open product development will require investment: It took nearly a decade to build out the GNU linux free software/open source operating system, and some years after that for open source applications like Firefox to emerge, and even more time for open source to become a standard form of software development.
Consumers and users: Consumers must create a healthy marketplace by choosing and rewarding products that protect them and the broader values of openness. We need to create a culture where consumers look for products that solve their problems, protect their privacy, or protect them from hateful content - and a market that encourages a diversity of products that meet these needs to exist and succeed. Consumers should demand algorithmic transparency and accountability to support a better market. Just as we ask each other if we’ve seen good films, or read any good new books, we should ask each other what new products we’ve seen that improve online life. And just as “shop local” and other consumer movements have helped reinforce the positivity of local retail establishments against the dominance of centralized conglomerates, so would “download open” or “install healthy” help promote a better software ecosystem.
Regulation: The idea that regulation would necessarily be a roadblock to open values is giving way to the recognition that guardrails are essential for online innovation to develop in the interests of people. We need to seek out smart regulation that is consistent with and reinforces open values. New data protection laws, like the GDPR, stand out as useful examples. While privacy and openness are often portrayed in conflict, today’s data protection laws are largely demonstrating how privacy protections can further open values of access, empowerment and opportunity.
Governance mechanisms: We still need healthy governance systems, and as a start we need to reaffirm and strengthen multi-stakeholder processes. Some efforts have been underway to do so. For some, the multi-stakeholder process is seen as yet another way for large (mostly US) companies to insert themselves in the discussion and use their wealth and power to lobby for solutions that reinforce their position. Nevertheless, multi-stakeholder processes continue to be a powerful tool to address a range of issues.
Case Study
Reimagine Open in Practice – Digital ID: Digital ID systems are increasingly impacting privacy, security, competition, and social inclusion. Identity is almost always mediating our interactions online, from corporate giants like Apple ID and Facebook login to government IDs which are increasingly required to vote, access welfare benefits, pay taxes, or receive medical care.
The choices governments make in designing and operating digital ID systems will determine if an ID system will be empowering or exploitative and exclusionary. Mozilla’s recent digital ID white paper shows how openness provides a useful framework to guide and critique these choices. We demonstrate how five elements of openness – multiplicity of choices, decentralization, accountability, inclusion, and participation – can ensure that identity systems put people first.
Section IV Conclusion
Open internet designers, activists, pioneers and advocates have undoubtedly made mistakes in pursuit of a connected world. As a community, we may have been slow to respond to the accountability problems in today’s internet due to our love of the values originally enshrined in “open” and “the open internet.” We should examine this history closely.
But the broader Internet community need not break faith with the power of openness principles to create the connected Internet we want. We must squarely face the problems of online life today, and contribute to developing solutions. We cannot cling to the original practices for fear that something worse will develop. As we do this we must equally hold tightly the key concepts that were originally captured in “open” and ”the open internet” and made it such a powerful and successful force.
To build the online life we want, the Internet community itself must be the voice for these values. We must build these values into new products and network designs. We must provide models that policymakers can reference to develop workable, measured, effective regulations, and we must give consumers and citizens ways to make meaningful choices for a better future. We must continue to build an internet and internet life that encourages broad-based empowerment, opportunity and participation as well as human decency and civility.