The European Commission Directorate General for Communications Networks, Content & Technology, a.k.a DG Connect recently launched a survey (deadline April 10th 2016) on the ‘future of the internet’ as part of its Net Futures agenda, which was established to “pioneer and coordinate research, innovation, and policy initiatives on what lies beyond the current internet architecture, software and services.” Below is a copy of my submission to the survey.
Status of the Internet in 2016
What is your assessment of the current status of the Internet and its impact from a European perspective? What are the major trends? How important are issues like Net Neutrality, de facto standardisation, broadband access, European Research Networks? What are the major trends in Internet usage and their impact on infrastructure and business models?
Net Neutrality and standardisation are hugely important requirements for the internet as a force to support the EU values of democracy and economic growth that supports all sectors of society.
European Research Networks (like the Network of Excellence in InterNet Science (EINS)) are important in stimulating development of research cooperation across the EU especially for better understanding of the current and potential future social impact of the internet, and developments like social media platforms, on citizens and communities. European Research Networks and projects like D-Cent are vital if the potential of the internet as a means for increasing citizen participation in civic and political life, at all levels (local, national and EU), is to be realized. Citizen led peer-2-peer communities and open source projects are of great value for producing new tools for the internet, but anything that requires larger system level or infrastructure support, including policy engagement, will need the institutional support and expertise that Universities and research institutions bring. Corporations can play an important supporting role in this, but are unlikely to take the lead, due to their fiscal obligations that focus their efforts towards short/medium term return on investment.
There is an undeniable trend among young people towards the use of mobile devices, especially smart phones, for accessing social network sites and engaging in social communications. The ‘app’ based model of internet interaction on these devices is one of the factors contributing to the growing fragmentation of the user experience of the internet. Apps produce a siloed, or ‘walled garden’, experience of access to the internet, both in term of the ‘look-and-feel’ of the interaction and in terms of the ease of exploration beyond the boundaries of the app’s platform.
Regardless of minor variations in implementation, the dominant business model for internet services remains one of capitalizing on user generated data, mostly via advertising, which is driving the push towards service personalization as a means for justifying the need for ever greater levels of user data harvesting.
View towards 2025 and beyond
How do you think the internet will look like in 2025 and beyond?
What are the major technology barriers in terms of open, generalized access to information? What mechanisms could be envisioned to help find the right information (including from sensors and actuators – IoT) when you need it?
What are the major issues and challenges in terms of application development and subsequent scaling-up in view towards generating global players? How could this be addressed beyond providing appropriate financing?
Prioritize the major research focus areas identified above in the short (18-20), medium (20-25) and long-term (>25). Highlight any potential disruptive areas.
The main developments that are shaping the internet in the near future (2018-2020) are:
1) The growing role of algorithms (increasingly partly Machine Learning based) as intermediaries that guide/filter/recommend or otherwise mediate the information, communication and services that people experience through online platforms. This has the potential of producing challenges to the core foundations of our democratic values of equality and freedom from discrimination. If algorithmic personalization results in platforms providing people with different services based on profiles that have been constructed about them, how can we safeguard against unjust discrimination (especially when Machine Learning means the algorithm’s behaviour can adapt in ways that were not anticipated by the platform designers)? At the same time, the rapid growth of online information is making it increasingly difficult for people to manage the data flows (e.g. social media data) without some form of algorithmic support. The challenge is therefore to develop methods for auditing algorithm behaviour and providing transparency about the algorithms to users so that they can make informed choices that will allow them to retain agency/control over their own online behaviour and experiences.
2) Cybersecurity will grow increasingly more important to the extent that the needs to protect against hacking, data thefts and other forms of cybercrime will mandate pervasive use of strong encryption. Since groups that pose national security threats will increasing shift towards using encryption anyway, the risks to online services when not using encryption will outweigh the potential benefits to security services. In the medium term (2020-2025) there will be a strong push for the use of homomorphic encryption that can allow computations to be carried out on data without the risk of decrypting.
3) The introduction of the GDPR and the general auditing and revisions of data management structures of online services provides an excellent opportunity for the EU to take the lead in developing a new, citizen centred, approach to online service provision in the near future (2018-2020). Instead of continuing the currently dominant paradigm of commoditization of personal data and user generated content, the EU should explore alternative technological and business model frameworks like the Nordic MyData model that rebalance the power distribution between service users and service providers. Such a rebalancing is required to better respect the content production that users are contributing to online platforms (e.g. social networks) and thus gain enthusiastic user engagement.
4) In order to provide true service quality based competition and counteract the ‘lock-in’ on social networks that is cause by the ‘network effects’ (i.e. people stay on a social network not because of the quality of the service, but because their social circle is on that network) there is a need to develop universal API standards for social media to facilitate interoperability between social networks. Standardized information transfer APIs that facilitate full interoperability between platforms would allow users to choose the service that gives them the functionality they want without losing the networks they interact with. The brand of your device should not limit the social circles you interact with. The application of data standards to improve interaction between applications/online platforms would provide a non-monopolistic plug-and-play ecosystem that generates cooperative network effects between service providers to let them form alliances that are capable of competing with the large, but siloed, platforms that currently dominate the internet.
5) The ‘Right to be forgotten’, the growing popularity (especially among young people) of services like Snapchat that promise to show messages only for a short period of time, and the endless ‘digital skills’ lessons about the dangers of posting embarrassing things online that could in the future harm their chances of getting employed, all of these highlight the social importance of transitory communication. The ability to forget and to experiment without having to fear that embarrassments of mistakes will haunt us for the rest of our lives is vital good that is necessary for society to function. Unfortunately, the way in which data is stored, copied and transferred online is not compatible with this, a problem that is massively magnified by increasingly powerful search engines. To cope with this poses the interesting challenge of developing ‘limited life-span data structures’ that delete themselves after a predetermined amount of time, or interactions, even when they have been shared onto different devices.