2018 BPF on Internet of Things (IoT), Big Data, Artificial Intelligence (AI)

 

Final output

Download .pdf

 

Executive Summary

The IGF Best Practice Forums (BPFs) bring experts and stakeholder together to exchange and discuss experiences in addressing Internet policy related issues.

 

The BPF on Internet of Things (IoT), Big Data, and Artificial Intelligence (AI) was part of the IGF intersessional work programme leading into the 2018 annual meeting of the Internet Governance Forum (IGF) in Paris, France, on 12-14 November 2018. This report reflects the work of the BPF and is the result of a community-driven bottom-up and open process.

 

Devices, networks and applications used by billions of users around the world generate a vast variety and high volume of data. IoT, Big Data and AI play a critical role in connecting, analysing, and generating value from this growing amount of information. IoT, Big Data and AI discussions are present in many fields, in the on- and offline world. The BPF’s focus is on where these technologies are used in consort in an Internet context and where Internet governance can play are role in stimulating further development and widespread use, as well as helping to avoid unintended negative side-effects.

 

The expectations on how IoT, Big Data and AI are going to contribute to solving complex problems and facing global challenges related to the environment, transportation, health, etc. are high and complement a fast growing list of examples of how they support individual Internet users’ daily lives.  There are however a large number of unknowns, potential impacts, risks, and social and economic implications that ask for guidance,  measures and policies for managing the impact of applying IoT, Big Data, AI technologies.

 

Stakeholder dialogue is crucial to allow the Internet to embrace IoT, Big Data and AI to the benefit of all. Each stakeholder group offers a unique understanding of how these technologies impact daily life, how to balance innovation with potential risks, and how to make the best of the opportunities while seeking ways to mitigate unwanted side-effects .

 

The BPF identified existing platforms and communication mechanisms for stakeholder discussion and collected examples and good practices of how stakeholder cooperation can help to media problems, avoid issues and support the use of IoT, Big Data, AI in the Internet. They are listed in section IV of this report.

 

 

The BPF suggested two Best Practices regarding definitios of IoT, Big Data, and AI that could contribute to an efficient and effective stakeholder dialogue dealing with the IoT, Big Data, and AI in an Internet context:

 

#1 - Define your terms narrowly so that it is clear for policy makers and stakeholders what aspects of these technologies they are discussing.

Not doing so can lead to sweeping generalisations or proposals that are meant to address a problem with a narrow technology or specific application that could have a range of unintended consequences. Worse, conflating different technologies and different applications will cause discussions to lose focus and is likely to create fear.

 

#2 - Be ecumentical about technology (or “Strive to be technology-neutral”).  Because technologies are changing so quickly and because potential problems with a specific application of a technology may or may not develop (or may be solved rapidly), it is dangerous and unproductive to try to write laws and regulations that cover one specific type of technology or one specific type of application.  Best practices should focus on what an application DOES not on how the technology DOES IT.

 

 

IoT, Big Data, AI are powerful technologies and when combined they become even more powerful tools that can be used for good or evil. Using IoT, Big Data, AI in an Internet context, creates a number of Internet Governance challenges. The BPF discussion pointed amongst other to the cybersecurity of IoT devices, risks related to AI mass data processing, and potential threats to human rights, security and social cohesion. The BPF had extensive discussions about the necessity that they are and how these technologies could be applied and further developed in ways that reflect ethical considerations and human rights.

 

BPF identified Best Practices for stakeholder to take into account when discussing the use of IoT, Big Data and AI:

 

#3 - Collaborate to ensure that these technologies are deployed in ways that protect user privacy and security, and network resiliency while fostering innovation. Stakeholders should communicate openly about the impact new technologies have on the public and existing networks and find ways to work together to develop future-looking policies.

 

#4 - Consider ethics and human rights when applying IoT, Big Data, and AI from the outset in the development, deployment and use phases of the life cycle.  This requires that users are aware of the benefits and risks deriving from these technologies.

 

#5 - Watch out for bias and incomplete data sets that may reflect only a small subset of the “real world” due to the Digital Divide, due to national regulations that restrict the export of consumer data, due to marketing decisions to only focus on certain geographies, demographics, or industry sectors.  In some cases, statistical techniques can weight data to compensate for some problems.  But in ALL cases, the limits of the data and Big Data analysis should be recognized.

 

#6 - Make privacy and transparency a policy goal and a business practice. Potential problems must be recognized before they become serious. Transparency is one of the most effective ways to nurture trust, and can for example be achieved by the publication of transparency reports and such reports are likely to become more common and more detailed as the IoT enables data collection about more intimate aspects of our lives.

 

#7 - Ensure systems are adequately secured before they get to the market. A balance will need to be found to distinguish “flaws resulting from irresponsible behavior” to flaws that could not be foreseen at the time, whereas system development has followed good practice - industry self-regulation may be the best way forward as to avoid regulation that is stalling innovation.

 

#8 - Foster technologies and business practices that empower SMEs. The growth of edge computing and “serverless computing” promises to give SMEs much cheaper and simpler ways to create the software needed to exploit the power of the data generated by the Internet of Things.  The best response to the threat of “Data Dominance” is not regulating monopolies, it is ensuring their are not monopolies by ensuring vibrant competition.

 

url 2018 BPF IoT, Big Data, AI Final Output:

https://www.intgovforum.org/filedepot_download/6733/1438

 

 

 

The BPF IoT, Big Data, AI is an IGF intersessional activity that aims at stimulating discussion and best practice exchange among Internet Governance stakeholders. The BPF is an open platform and all interested are warmly encouraged to join by subscribing to the mailing list and participating to the BPF's meetings.

Draft Output of the BPF IoT, Big Data, AI

The BPF IoT, Big Data, AI has published its draft output report.

  • download the report here (.pdf)
  • give your feedback on the draft report here 

 

 

Call for contributions  (closed)

The Best Practice Forum on IoT, Big Data, AI wants to collect best practice examples of how Internet Governance and stakeholder cooperation can help to give guidance and/or to develop "principles" and as such can help to stimulate the further development and application of IoT, Big Data, AI in an ethical and sustainable way. 

The BPF wants to highlight success stories and good practices, but also learn from experiences and pitfalls. 

Contributions can be submitted to [email protected] , preferably as a word document, and will be reflected in the BPF output document.  (please submit before 15 October)

The following questions are intended as a suggestion to help to structure your case study:

  1. Please provide a short description of the practice, experience or example. (1/2 page max)
  2. Please indicate why it is a good or bad practice, why it was successful or not.
  3. What are lessons learned from the example?
  4. Where to find additional information (please provide links, docs, etc. )

 

Mailing List Sign-up

The BPF IoT, Big Data, AI mailing list is open to all stakeholders interested in or with expertise on related issues. Sign-up at https://intgovforum.org/mailman/listinfo/aiiotbd_intgovforum.org

BPF IoT, Big Data, AI Meetings 
Thursday 18 Oct, 12 pm UTC - Virtual meeting IX - meeting details
Tuesday 30 Oct, 12 pm UTC - Virtual meeting X - meeting details
Thursday 8 Nov, 12 pm UTC - Virtual meeting XI - meeting details
Wednesday 14 Nov - BPF workshop at the IGF 2018 - meeting details

past meetings 

Background

The application of Artificial Intelligence technologies like speech recognition, chatbox, etc. to internet are considerably increasing, significantly modifying our way of life. At the same time smart sensors distributed to several objects collect and distribute a huge amount of data over the Cloud.

Thanks to the availability of big data and cloud computing, main players like Google, Facebook, Amazon can classify and group user data by means of machine learning applications, to extract user profiles and models and apply them in a variety of contexts. Self-learning algorithms are also available on our mobile devices, influencing our choices, our purchases, etc.

BPF proposal on AI, IOT and Big Data Analysis in Distributed Cloud Environment is specifically targeting a study group environment where IGF may discuss on these buzz words holistically and thereafter the output of this process will result in the development of a document on these topics. This may help to take maximum advantages of the technologies and to avoid any misuse that may be harmful to person, society or nations.

The proposed BPF is very much relevant with these days’ multistakeholder IGF discussions. It will take advantage and will analize the main results that have already discussed in different working groups, forums and internet community i.e. ICANN, RIRs, IETF, ITU, AgID, IEE, OECD, etc. organizations that are already debating those technologies with their own communities.

There are several benefits coming from AI, IoT and Big data: they can help improve efficiency and cost saving; they can enable better and more accurate decision-making and can be used for discovering specific data patterns, complex cause and effect relationships, etc.

This could help people to solve complex problems and global challenges related to the environment, transportation, health, etc. Also on the level of the individual internet user there is an almost endless and growing list of examples of how AI could help. There are AI applications, for instance, designed to support doctors in diagnosing and detecting symptoms of diseases like Alzheimer and Parkinson, applications supporting patients or helping to improve the lives of people with disabilities, and AI in imbedded in tools intend to assist people in organising their daily lives and work. Furthermore, AI supported BIG Data Analysis on human genome data may lead to early diagnosis to some diseases Diabetes. On the other hand this may lead to develop biological weapon against a particular race.

While those new technologies could help to tackle many global challenges and support people in their daily pursuits, there are a large number of unknown and potential negative impacts and risks, for example related to privacy and discrimination, inequality and infringement of human rights.

This is an issue of importance to all stakeholder groups, as all will be impacted by and benefit from them.

It is important establish a dialogue with all stakeholders to understand how Internet could best benefit from AI, IoT and Big Data and to identify the risks and seeks ways to mitigate unwanted side-effects, e.g. by setting limits to the automated decision making.

Several social and economic issues related to the application of the AI, IoT and Big Data to the internet can be identified. Therefore, a debate would be required to help stakeholders to pick out the right measures and policies to be adopted for managing the impact of those new technologies on the society.

For AI, for instance, it would be important to request transparency in the AI algorithms, as they may induce decisions with severe impacts on people, they can categorize people based on their preferences, etc. AI applications should be designed to respect common society rules, cultural and religious difference, etc. (AI Ethics).

For Big Data Analysis in Distributed Cloud Environment is particularly very important to mutually discuss in a multi-stakeholder platform regarding the distributed cloud environment services where data analysis are used. Because, block-chain alike services and products will start flooding our world sooner than we are expecting. Moreover, lifestyle sharing services like Uber, AirBnB, etc. is having a parallel domain than the conventional services.

 

Objectives

BPF proposal on AI, IOT and Big Data should specifically target to be a multi-stakeholder and multi-disciplinary platform to understand the wider context of ‘Artificial Intelligence’, ‘IOT’ and ‘Big Data analysis and distributed cloud’ for each stakeholder group. Moreover it should discuss on opportunities and threats related to the application of these technologies on Internet, and ways to stimulate the positive or mitigate the negative. Focusing on the benefits, the use of the AI, IOT and Big Data could be analysed for reaching SDG objectives.

The BPF may be started with the study of the context of Artificial Intelligence, IoT and Big Data. Several focus groups may be organized, one for each of the following tasks:

1. Identify existing platforms and communication mechanisms between stakeholder groups discussing or well-placed to discuss AI, IOT and Big Data issues;

2. Analyse the main AI, IOT and Big Data applications used in Internet and the related benefits/threats;

3. Identify common problems and best practices to fix them;

4. Activate a network with the main existing working group on AI, IoT and Big Data (ICANN, RIRs, IETF, ITU, AgID, OECD, IEEE, etc.) to share results and best practices;

5. Identify how AI, IOT and Big Data can be used to reach SDG objectives;

6. Identify the impacts on policies and regulations as well as budget implications for governments (computing power, bandwidth, technical capacity);

7. Verify roles and responsibilities of the different stakeholder groups and explore if it there is ground for global common policy recommendations as outcome of the BPF.

During the first virtual meeting the BPF will decide how to distribute the tasks for 2018 and 2019. It could be reasonable to have the first 4 groups of activities managed during the 2018 leaving the others to be developed during 2019.

 

Expected Output(s)

As anticipated before, the BPF intends to stimulate the debate between stakeholders, laying the basis for a continuous dialogue, within and outside the context of the IGF. A tangible output of the process will be one or more documents, based on stakeholder input, on the benefit and threats related to AI, IOT and Big Data and the main results reached by the existing forums.

Those documents could help to take maximum advantages of these technologies and to avoid any misuse.

An initial document will be issued in 2018 that should report the main AI, IOT and Big Data applications used in Internet, the related benefits and the main results and best practises coming from existing working groups (ICANN, RIRs, IETF, ITU, AgID, OECD, IEEE, etc.).

 

Outreach and Dissemination Plan/Strategy

As mentioned, we envisage a broad participation from the different stakeholder groups and multi-disciplinary input, and intend to obtain this by a mix of general outreach to all interested parties and targeted outreach to existing working groups on AI related issues (e.g. ICANN, RIRs, IETF, ITU, AgID, OECD, IEEE, etc.).

AI, IoT, Big Data and Distributed Cloud Environment is having their transmission and delivery platform based on internet. That’s why IGF has a much greater role to play to have policy level alignment among economies before the flood gate of these high computational services are opened.

The main results discussed during the first year by the BPF will be collected in the first report that will be presented during IGF 2018. We expect that the different stakeholders involved in the BPF’s discussions will take the BPF’s report back to their respective communities. A second report will be issued on 2019 that will be more focused on how AI, IOT and Big Data can be used to reach SDG objectives and on the impacts on policies and regulations.

 

Documents 

Proposal to the MAG for 2018 Work