Roberto Viola, Director-General from Communications Networks, Content & Technology at the European Commission speaks about REIsearch and the Internet of Humans: how we would like the internet of the future to be

Check against delivery

I am very honoured to be the 2017 Bruno Kessler lecturer and I am also humbled to be following previous talks held in former years by world famous international scholars like Jared Diamond, Marcus du Sautoy, and Heiner Bielefeldt.

I am delighted to address in this lecture a topic that is very important to me and I think to you all: the Internet; the great opportunities it provides, the concerns it raises and how we as Europeans can contribute to the development of a more human-centric Internet.

If we go back in time, it was in 1991 when CERN released their first web browser to the general public; this made the Internet accessible outside of the scientific community. This was a turning point in the evolution of the Internet; in a little more than 25 years there has been an unprecedented expansion in terms of use and available services. In this period, the number of users has grown exponentially, from a few millions to several billions nowadays. Today the number of Facebook users is approximately 2 billion, the number of websites reached is a bit less than a billion. Every minute the Internet handles approximately 400.000 new tweets; 3.5 million google searches; 200 million of YouTube videos are watched, and 260 million emails are exchanged.

Exponential growth has made the Internet a major economic driver. It is difficult to estimate what  the size of the Internet market is, but more importantly the Internet enables digital transformation in most sectors of the economy, thus challenging the business models of many traditional sectors.

But this development is not only about numbers and money.

What is more important is how the Internet transforms our society: the way we live, the way we work, how we get informed and relate to each other. And let’s not forget its international dimension; borders drawn by humans are disappearing in the digital world, when we exchange goods or interact with each other. For many of us our normal daily life activities are now unpredictable without the Internet. How many times a day do you check for facts or look for timely information on the web? Can you manage your personal relations without e-mails, instant messaging or social media? Can you plan a trip without Internet, to check for flights, hotels, or places to visit?

For better or worse – this is the world we live in.

The growing importance of the Internet implies that it is no longer just an IT network of networks: it is rapidly shaping the economy and our daily lives. This raises fundamental questions. Is the current Internet responding to the needs of the citizens? What kind of Internet do we want in ten-twenty years from now? What is the role of policy-making in driving the evolution of the Internet? What is the role of the civil society? What ground-breaking research we need? Can we leave the innovation in the hands of few or there should be a bottom-up democratic participation to shape the future?

Today, still many Europeans still hesitate about going on-line and have concerns about personal data and transparency. For example, last month, the EU-funded project REIsearch with an international team of journalists and researchers have tried to better understand what EU citizens think, feel, fear, and express about future Internet technologies and their potential impact on society. A network analysis was performed by investigating – via machine learning and natural language processing techniques – over 650,000 messages around the future of the Internet, those messages were generated by 350.000 users between November 2016 and April 2017, in 54 languages, on Twitter, Facebook and Instagram. Just to share a few of their main findings:

  • Fintech: discussion threads are systematically addressing the financial implications of Internet technologies and how new networked technologies may open up novel and disruptive financial scenarios, including digital currencies and blockchains;
  • Cybersecurity and Cybercrime are among the most interconnected topics found in research, meaning that concerns about them are present in a large part of the discussions;
  • Privacy is among the principal concerns expressed about the development of the Internet;
  • Algorithms is another one, both in a positive sense (for example as a driver for applications in health and smart systems) and in a negative one (for example due to possible negative implications in terms of privacy, control and surveillance);

REIsearch also ran a communication campaign with the support from nine EU newspapers, including Il Sole24Ore in Italy, publishing three articles on the Next Generation Internet for three consecutive weeks, and asking their readers to complete a survey and voice their opinions on these topics. Nearly 8,000 people took part in the survey, and about half of them completed the questionnaires. While this is not a scientific survey and the sample is not representative of the entire EU population, the initial findings are interesting:

  • Automation of jobs is a source of great concern, and people expect public institutions to manage the disruption caused by rapid technological advance.
  • Internet technologies are seen as a cause of rising populism, but also as an opportunity to restore public trust, well informed debate and multi-sectoral collaboration.
  • survey participants were sceptical that Internet technologies can tackle financial, social, or economic inequality (12-25% strongly disagreed) they strongly agreed that future Internet technologies could improve our society through increasing access, availability, cost-effectiveness, and personalisation of public services (24-38% strongly agreed).

There are also concerns on the way the Internet market might develop. Fuelled by network effects, we have seen the emergence of large platforms. From a user’s point of view, the Internet should not be a set of silos, of non-interoperable, proprietary services.

When policy-makers, researchers and society reflect on the future evolution of the Internet, we should take a fresh look at all these issues. The Internet should offer more to the people and to our society, providing better services and greater involvement and participation. It should be designed for humans, so that it can meet its full potential for society and economy and reflect the social and ethical values that we enjoy in our societies.

Towards a Next Generation Internet

What is my vision of the key technology building blocks that will shape the Internet of the future? Seamless interaction between real and virtual world, cooperation between man and artificial intelligent agents will shape the future providing for example better health care and safer and cleaner transport systems. However to unlock this potential trust by users is the precondition. Universal identity, personal data spaces, and security of transactions are expected to be major enablers to build this trust. In the rest of my talk, I want to focus on a number of key technologies that in my opinion constitute the foundation of the internet of the future without having the ambition that this list is exhaustive.

Artificial Intelligence

Recent advances in Artificial Intelligence will be critical to turn the increasingly growing mass of information into knowledge and to provide autonomy and intelligence into networks, robots and connected objects. Indeed, a next wave of innovation will come from our seamless immersion in increasingly smarter physical and virtual environments powered by Artificial Intelligence. The potential is enormous to assist people, improve their lives and also for the society at large (I already mentioned healthcare and safe transport but also better education, more sustainable food production, security of our society, more resilient financial systems).

Artificial Intelligence and robotics are key drivers of future economic and productivity growth. Many sectors including health and agriculture will benefit from these technologies, be it faster and more accurate diagnosis of diseases or a more favourable environmental footprint that can reduce pesticide usage. The economic impact of the automation of knowledge work, robots and autonomous vehicles is estimated to reach between €6.5 trillion and €12 trillion annually by 2025.

To seize these opportunities in Europe, we have set up SPARC, the Public-Private Partnership for robotics in Europe. With €700 million EU funding and, adding the private investment, an overall investment of €2.8 billion, SPARC is the biggest civilian research programme in this area in the world. In the period from 2018 to 2020, we will invest an additional €50 million to build a European Artificial Intelligence-on-demand platform. Our vision is that of a platform where every potential user of Artificial Intelligence gets help and support to implement the latest algorithms in their products and services. Developers will learn about user requirements, and users will understand better what current Artificial Intelligence algorithms can do for them and their business.

Collaboration and openness will be key to achieve these goals, and we will make sure that these principles are deeply embedded in our work on Artificial Intelligence. We envision a platform where knowledge is openly shared and accessible. Artificial intelligence has to be democratised and made available as widely as possible if we want to reap its full benefits; and this is what we are aiming at with our AI-on-demand platform, an open ecosystem where AI technologies and applications can develop.

But with the increasing level of autonomy of systems and algorithms, there are also increasing concerns. It is essential for Europe to make the necessary efforts to lead the Artificial Intelligence revolution, to ensure that it will be respectful of and driven by the core European values that characterise our society.

In particular, there are concerns about the transparency of algorithms and decision making processes. It is therefore critical to investigate solutions capable of explaining these decisions, which may sound as paradox to you but it is scientifically challenging for some of the AI-based approaches, and requires more research into explainable Artificial Intelligence. I think that seeking for openness, transparency and predictability is scientifically sound and pragmatic approach to the complexity of the debate on accountability and liability of AI systems.

What about Internet of Things and AI In this case, it is not only about extracting information from data collected by sensors and devices, but how to use such information to increase the level of autonomy of the systems around us, such as cars, robots or connected devices, and make them more useful to their users.

One example is Digital Industrial Platforms that bring together different technologies, applications and services. They open up data from the machines, products and operators to make them accessible to monitor processes, to connect different stakeholders, such as users and application developers. Good examples are platforms such as Industrial Data Space (IDS) and reference architectures such as Reference Architecture Model Industry 4.0 (RAMI). A reference architecture gives a framework to position different applications, specifications, standards with respect to each other, and promotes common understanding.

I believe that public funding is instrumental to support these developments in areas such as smart connected factory clusters that include SME’s, smart hospital and care, smart agriculture, connected and automated mobility, or smart energy.

The Internet of Things can be as simple as heating your home efficiently or as complicated as precisely monitoring pollution levels in the environment: either way, billions of devices transmitting an ever increasing number of data will drive opportunities but also complexity. This raises new challenges for improved methods of search. Not only what we today search in internet shall be discoverable but this shall extend also to objects and elements of the physical world. One example is cognitive search that takes into account contextually relevant information and adds intelligence to the human-machine communications. The notion of relevance is simple for humans, but it is less obvious for search engines to take into account in order to improve situation awareness and decision making. Cognitive search combines technologies for natural language processing, semantic analysis and data visualisation, opening promising paths to new forms of information access. Once again semantic algorithms will be effective if the language to describe objects is shared in widely accepted standards. Would you confidently enter a room where your senses cannot acquire information about the environment? The same applies to a complex cyberspace, we need to know how to name objects and what they are. Semantics in internet needs a common and shared effort.

At the same time, new legitimate concerns are raised regarding security, privacy, liability, ethics and acceptability in this new generation of connected objects, which will handle important amounts of sensitive data. For example: in case of technology failure or security breach, there may be some physical risks for patients using connected objects to monitor their health or well-being.

The deployment of the Internet of Things might create scenarios which go beyond the scope of existing legislation. We need to develop a framework that facilitates citizen acceptance, notably in terms of security and trust.

Interactive Technologies

Another area where we are witnessing a technological revolution is Man-Machine Interactive Technologies that is producing a paradigm shift in the way humans interact. Speech Recognition and Augmented Reality are two of the most noteworthy examples of this evolution and are uncontested game changers. One out of five search queries on mobile phones is already speech-based in the US. And this figure is expected to rise up to 50% by 2020. By 2019, the voice recognition market will be a 600 million € industry while the latest figures for Augmented and Virtual Reality are around 120 billion € in 2020.

Interactive technologies are delivering information in more natural, efficient and less intrusive ways, providing enhanced and personalised user experiences. In Augmented Reality, Pokemon Go hit the market less than a year ago, it was downloaded more than 650 million times, gained 600 million € within 90 days and still have a daily revenue of 2 million €. In Virtual Reality we have seen recently the first general public headsets hitting the market with consumer applications that are so far mainly limited to the entertainment area.

The technology is also slowly penetrating traditional markets. For instance, the IKEA application allows the user to see the furniture virtually placed in their house or the applications used by architects or kitchen makers to better illustrate the foreseen modifications in a house. These types of applications common in the Next Generation Internet and will impact a large variety of industries such as the education, health, tourism, cultural and creative industries.

While still too cumbersome and bulky, current VR interactive systems (think how they will be considered as funny in 50 years from now) will start delivering soon into single-user experiences. However once the technology moves to a further step in the direction of usability, there are enormous unexploited opportunities in multi-user social interactions, for example in virtual collaboration and co-creation. The team experience is the next breakthrough with far-reaching market opportunities but also social implications. And this requires a combination of competences and technologies that can be referred to the Next Generation Internet:

  • hardware and software to provide more realistic and natural experiences, including a larger field of view, light field, panoptic capture, focus free, photo-realistic rendering, increased resolutions or frame rates;
  • research on social interactions to develop theories and technologies allowing an augmented human experience through technologies such as augmented reality, virtual reality or brain interface, to interact, work or entertain in groups, thus developing new ways of social interactions;
  • support the transfer of these technologies across different sectors (industrial manufacturing, automotive, data life cycle, consumer goods, healthcare, public services, design, entertainment, media, culture…)

A human centric Next Generation Internet shall reflect the openness, diversity and the inclusion that are at the core of European values. We want an open Internet that allows every citizen to interact and, from all walks of life, to take part in the online society. We want an internet that empowers citizens not discriminates them. Let me mention two very concrete hurdles faced by citizens today every time they go on the Internet:

Multilingualism represents one of the greatest assets in terms of cultural diversity in Europe and, at the same time, one of the most substantial challenges for the creation of a truly integrated EU. In the EU, there are 24 official languages and more than 60 national and regional languages. A recent study by the European Parliament shows, however, that:

  • out of the 66 % of public administration portals that offer information in languages different from the country’s official language, only 39 % offer information in a language other than English;
  • 90% of European consumers prefer to browse websites in their own language and;
  • 42% never purchase online products and services in languages other than their own.

These numbers are telling. How can we equip our citizens to make the most of digital technologies with half of them reluctant to purchase online in a different language than their mother tongue?

Today, Language Technologies are increasingly embedded in mobile communications, in intelligent assistants. They have allowed a multitude of innovative IT products and services in wide range of industries. The use of Language Technologies in the automotive sector for example will witness considerable growth as more and more (connected) cars are being fitted with voice-activated controls of various kinds. We are only at the beginning of an ever more, intelligent, deeper human-machine interaction.

There are already virtual assistants who recognize speech queries and answer verbally, such as Siri, Cortana and Google Now. They are relatively primitive compared to humans but start to work efficiently for routine tasks in the main languages, while even for basic interaction they still lack quality for many of our European languages.

The challenge will consist in ensuring that the Next Generation Internet is “language transparent”, that devices and machines are able to fully understand various languages and interact deeply with humans. It means that the research and innovation we will support will have to bring together language technologies, computational linguistics, artificial intelligence, cognitive science, neuroscience to name a few, to tackle hard core research and innovation issues on the meaning of human language and on language technology components for seamless communication across languages and sectors.

The second core value is inclusion. The Internet has been, justifiably, regarded as having enormous potential for promoting the inclusion of people with disabilities. Taking into account demographic ageing, it is expected that in 2020 approximately 120 million persons in the European Union will have multiple and/or minor disabilities. Today’s 80 million do not have full access to available products and services, including public web services. The challenge of the Next generation Internet will be to serve the needs of these consumers. I cannot envisage an Internet where two thirds of European citizens will be able to interact seamlessly with personal assistants while one third will not have access to basic services. And this is very much a win-win situation as there is a demand for accessible products and services from an increasing number of citizens with disabilities and/or functional limitations, creating market opportunities for European industries.

Towards universal identity and personal data spaces

Today in Internet your identity is largely app-dependent and your privacy settings are app-dependent or platform-dependent. This is quite different from the physical world where you as an individual can decide on the level of interaction with others.

I also want to address other related aspects that will have a strong influence in the future developments of the Internet:

  • Privacy, security and identity management in the Internet;
  • Blockchain driving Next Generation Internet developments;

The fundamental rights to the protection of personal data and to the respect for private life (privacy) are reflected in the EU legislation, which is certainly the most advanced in the world: the General Data Protection Regulation and the rules on ePrivacy. These rules provide protection to individuals regarding their personal data and private sphere, establishing the conditions under which personal data can be processed. This means that processing of personal data must be lawfully and transparent; for a limited purpose, must comply with the principle of data minimisation and has to be implemented in a secure way. It is not needed that personal data is always anonymised to ensure the protection of the privacy of an individual. Safeguards are provided by EU legislation to ensure that processing of personal data can take place, while the rights of individuals are not compromised.

Consent is a corner-stone of the GDPR and the ePrivacy Directive (as well as in the future ePrivacy Regulation) for data processing , but individuals feel overwhelmed by consent choices or simply click away the requirement to agree to consent notices.

An emerging offer of personal information management services (PIMS), also referred to as “personal data spaces”, may be able to make consent work and thus give companies legal certainty on what they can do with the data. Companies that require good quality, up-to-date personal data for their operations and want to be sure that they comply with the GDPR may take PIMS as an opportunity. This includes online retailers, consumer brands, banks and insurance companies.

PIMS services can take many forms:

  • They can be comprehensive solutions offering a “personal data space” in which the individual stores his/her data and from which s/he shares such data with interested third parties.
  • They can be in-built features such as consent-assistants which help the individual to identify whether the terms and conditions of usage of personal data are compliant with his/her wishes.

For a number of years, living labs like the Joint Open Lab in which there is participation here in the Trento region, or the French MesInfos project have brought very valuable insights on user driven data spaces.

PIMS services still are currently in a state of infancy. They are essentially provided by start-ups that have a difficulty to create the relevant network effects that online platforms need to obtain to become relevant. The most important push from a regulatory perspective will be the upcoming GDPR with its reinforced requirements on consent, but most importantly the right to data portability.

With this effect, now we see that things are getting real:

  • UK start-up digi.me is partnering up with the Icelandic Directorate of Health, creating a living lab in which all Icelandic citizens will be able to download and port their health data through the digi.me app.
  • Start-ups receive seed funding from large players, including companies like Swiss Re.
  • Larger players are beginning to offer PIMS services, e.g. the recently launched AURA platform by Telefonica. Facebook collaborates with French start-up Chekk.me. Dt Telekom has just launched a competition on consent assistants

Along with management of personal information, there is the need for internet users to be identified regardless of the platform they use when and if needed. This brings up the issue of electronic ID standards – ideally leading to universal public eID. An ID that is platform independent, standardised and perhaps global. Identification of individuals is a key feature of citizen-centric society – the same applies online. Our European eIDAS framework, the EU framework on electronic identification and trust services, is a very good example how to secure universal eID and its use should proliferate as fast and as much as possible.

The use of a multitude of username and password combinations is both inconvenient and poses security risks. As a remedy, in order to keep identification simple and secure, consumers should be able to choose the credentials by which they want to identify or authenticate themselves. In particular, online platforms should be encouraged to accept credentials issued or be recognised by national public authorities, such as electronic or mobile IDs, national identity cards, or bank cards.

Europe wants to be a role model in the world to promote eID.

Security and the authentication of the communications are also critical for the Internet, given the ubiquity of electronic platforms and systems in today’s world. Cybersecurity, attacks, fishing and leaks are making the news regularly. As policy makers should look at these aspects very seriously and with a long term perspective.

Some of the best known encryption standards used worldwide were invented in Europe. But there is a constant race against hackers, so new techniques keep developing as for example homomorphic cryptography which allows to perform computations on encrypted information without decrypting it, making it more difficult for malware programmers to write viruses. This type of cryptography is many respects revolutionary, since it allows to search in encrypted data, or search and compute over encrypted databases without revealing any personal data. There are also many applications in the area of secure methods for cloud computing, distributed computing and delegation of computation in general, that can be implemented with homomorphic encryption schemes. Computation power is the challenge here but nothing is impossible in the coming era of high performance computing.

You probably know Quantum Technology. It is a very promising field that uses quantum effects of particles and their very specific properties. Some applications are all around us already for years, such as the lasers, magnetic resonance imaging, etc. But much more is to come as part of the so-called “2nd quantum revolution”.

One of the most promising is the realisation of quantum computers, which use the superposition principle of quantum physics to compute in a massively parallel way. This gives quantum computers an unprecedented power even compared to the best supercomputers. Such quantum computers are not there yet, but all the major labs worldwide are starting to build prototypes, and it can be expected that real quantum computers or hybrid machines will be with us within the two next decades.

This is wonderful for many scientific fields such as climate modelling, materials research, health, etc. But it poses a challenge to our cryptography, which relies on mathematical complexity, in the sense that you can’t hack them in any reasonable time, even with the most powerful computer. Quantum algorithms that can break today’s cryptography do exist, and are just waiting for a quantum computer to run.

So the cryptographers are now working on post-quantum crypto algorithms. Hopefully there is another approach that one can pursue in parallel: to use the quantum technology for cryptography! There are of course still some limitations so your mobile phone will not get that technology in the short term. But for specific communication with high security needs, commercial applications already exist. For example, the elections in Switzerland are protected via quantum communication networks. Quantum cryptography is a key technology to increase the security of our networks, to protect critical infrastructures, to authenticate and timestamp financial transactions, to ensure secure long-term storage of critical data.

Much has still to be done to ease usage, to provide plug-and-play devices that can easily integrate existing networks, and cover distances of thousands of kilometres. This is what we support in the Quantum Communication area of our FET Flagship initiative on Quantum Technologies. But the prospects are there, and Quantum Communications are expected to be a major contributor to the enhancement of the security of our communication networks.

Another technology I want to touch upon is blockchain as a driver for Next Generation Internet developments. Blockchain is, quite simply, a digital, decentralized ledger that keeps a record of all transactions that take place across a peer-to-peer network. The major innovation is that the technology allows market participants to transfer assets across the network in a verifiable and permanent way without the need for a centralized third party.

Blockchain was originally (and still is) the technology underpinning the crypto money Bitcoin. But there are now different types of blockchain: (i) eg public blockchain (like for Bitcoin or Ethereum), (ii) consortium blockchain (like the R3Corda in financial services, or Hyperledger from IBM), or (iii) private blockchain with many variations. And it is still continuously evolving.

The ledger itself can be programmed to trigger transactions or programmes automatically, taking into account external triggering conditions. With blockchain, we can imagine a world in which contracts are embedded in digital code and stored in transparent, shared databases, where they are protected from deletion, tampering, and revision. In this world every agreement, every process, every task, and every payment would have a digital record and signature that could be identified, validated, stored, and shared. Intermediaries might no longer be necessary. Individuals, organizations, machines, and algorithms would freely transact and interact with one another with little friction.

This is the immense potential and transformational nature of blockchain, that some considered as a revolution or a foundational technology like was the internet. Does it mean that blockchain is the future of Internet?  I am tempted to say that blockchain will not replace Internet as we know it but it would be an important part of its evolution. Blockchain applications are being built on top existing digital data, communication, and computation infrastructure. As we see trends today and while it is still in early stages, blockchain would be particularly helpful for tracing uses of data, transactions and products; guaranteeing executions; providing (digital) value to projects and behaviours; disintermediating or changing intermediation; creating trust. But all this, still often works at proof of concept or limited scale level, raising notably governance, cybersecurity/cryptography challenges, requiring progress also on interoperability and standardisation for easier integration in wider systems.

As part of the evolution of the Internet, blockchain through it various forms of design and exploitation, could become a key (and often invisible) component contributing to almost any socio-economic activity, a true value creator which reshapes economic and societal behaviours.

I cannot finish this flight over the next generation internet technologies without with some reflections on how to treat in large data sets and store them in an environment friendly way.

I want to address recent advances on data compression that are critical to handle large data sets (big data).

If it is true as is true that to achieve the next breakthroughs in medicine when need quickly a next generation of exascale supercomputers. To move large data sets around in the internet we need efficient data compression techniques. Traditional compression techniques are agnostic about the content of the data they are compressing in the sense that they are prepared to compress content of arbitrary nature, as represented in bytes. However, there are many domains (for example sound and image/video processing) for which more useful assumptions can be made that can be domain-specific. For example, it may be assumed that image files from a radiology lab will all contain X-ray images of features of human anatomy (as opposed to, say, static pictures from a TV set). When those conditions apply, advanced compression schemes have been very successfully applied that are based on a technique called autoencoding.

Images are first transformed to a simplified (and thus much compressed) representation, and then restored before use to a representation that is perceptually faithful to human observers (in the sense that humans are typically unable to tell the difference between the original and the restored representation). Based on these principles, Google has recently developed a machine learning data compression scheme called RAISR. This is reportedly saving up to 75% of bandwidth on images that need to be transmitted in response to requests for web pages.

The general lesson that these examples provide is that instead of having a few general purpose data compression schemes, we are probably moving towards a future where tens of thousands of data compression schemes will exist, each tuned and constantly updated to specific applications depending on the specific nature of the data to be compressed and the specific use that decompressed data will be put to. Once again sharing those compression standards in open format within the internet community is the way forward.

Large data sets need to be processed and stored. Eco-friendly data centres and data storage, are part of a more human centric vision of the future. In recent years, the world and the EU in particular have been deciding on major steps towards environmental and energy sustainability. The Paris Agreement signed in December 2015 continues to have full support of the EU. It sets challenging Sustainable Development Goals for 2020 to 2050. The emission of green-house gasses should drastically decrease, renewable energy increase up to 75 % and energy efficiency improve substantially, by approximately a factor of two.

European based data centres consumed more than 104TWh per annum in 2015, representing 3% of total electricity. This could grow by 20% by 2020, or 35% over 9 years. The estimated growth is even higher for networks: up to 150% in 9 years. Emerging and new ICT trends such as the anticipated explosive growth of Big Data and Cloud Computing, High Performance Computing, Internet of Things, Industry 4.0, etc. are not yet taken into account in the above growth figures.

ICT’s are enabling technologies for many economic, industrial and societal processes. This includes the increased use of ICT and the related energy requirements for reducing total energy consumption (e.g. using more ICT for saving energy in industrial processes, transport, smarter use of resources, etc). Nevertheless, the proliferation of ICT requires that ‘green ICT’ becomes an explicit requirement in future ICT developments. This includes more eco-friendly cloud and data storage, but also data processing and networking. Both energy production (low emissions, renewables, re-use) and efficiency of energy use (low power high performing processors, storage and networking, optimisation systems) need to be addressed, subject to ensuring that Europe can reap the continued benefits and functionality brought by ICT.

The European Commission has already dedicated substantial effort to this area, such as the Code of Conduct for Energy Efficiency in Data Centres. This initiative sets ambitious voluntary standards for companies willing to participate and over 130 companies have already signed up for it, representing 300 data centres based in Europe. We are in the process of preparing further initiatives to ensure a sustainable digital future. One of our priorities in our supercomputing programme is to develop a highly performant low power processor.

Achievements and future actions

What lies ahead of us?

In my vision of the Next Generation Internet, research has to lead technology development but it has also an important role in identifying emerging issues that will steer the policy debate. In the remaining of the Horizon 2020 research and innovation programme we proposed under the Next Generation Internet umbrella a set of technology areas that will drive the developments towards a more human-centric Internet: Artificial Intelligence, Internet of Things, interactive technologies, future media and social networks, as well as language technologies and inclusion. Universal identity and personal data spaces are also expected to play a major role. And one could foresee the Next Generation Internet continuing to be a very ambitious initiative also in the successor of the Horizon 2020 programme.

But more important than my vision, we need the collective vision and engagement of all stakeholders to create the Next Generation Internet. We need to gather and discuss ideas concerning concepts, technologies and applications coming from the largest possible community. We need to work together to build a future Internet that better responds to the expectations of the citizens. With the recent citizen consultations and NGI summit at the European Parliament, we have started a public debate that we want to continue and enlarge in the coming months. I encourage all of you to participate in the discussions and actively contribute to shape the Internet in a way that responds to your values and expectations.

https://ec.europa.eu/digital-single-market/en/news/internet-humans-how-we-would-internet-future-be