How the IGF has worked so far and improvements that could be made. Are there lessons for AI?
How has the IGF worked?
Originally proposed by the Working Group on Internet Governance, the Internet Governance Forum (‘IGF’) was formally called for during the second phase of the World Summit on the Information Society (‘WSIS-II’) held in Tunis. The resulting Tunis Agenda set out the mandate of the IGF, which was endorsed by the UN in April 2006 for a period of five years. This mandate has since been renewed twice, once for a further five-year period, and more recently in 2015 for a third period of ten years.
From its inception, the IGF was intended to be a multilateral, multi-stakeholder, democratic and transparent platform for discussion on internet governance issues. Its mandate included instructions to:
There does appear to be broad agreement that the multi-stakeholder aspect, and the emphasis that is placed on the importance of this, has been one of the major strengths of the IGF. Some have indicated that the emphasis on multi-stakeholder discussion has resulted in better quality decisions.
Inclusivity and facilitation of discourse
Others have opined that the IGF has effectively built a community of expertise across various different stakeholder groups, developing an inclusive culture with a ‘common language’, enabling more effective discourse and involving those who might otherwise not have been involved. The IGF has also encouraged the emergence of national and regional IGFs (known as NRIs), which has helped to strengthen the global framework for developing countries participating in internet governance at both the national and international level. This has in turn helped facilitate discourse between disparate groups including governments, intergovernmental organisations, private companies, the technical community, and civil society organisations, on a range of topics, including: proposed regulatory frameworks, potential risks, global trends and best and worst practices that have been adopted, or are currently under discussion.
In a 2013 article, Dmitry Epstein claims there are “a growing number of examples where one can trace the influence of IGF processes”; he gives the example of the 2010 UN Economic and Social Council where the working group initially contained only government representatives. A quick response by civil society groups and other non-state actors ensured that representatives from multiple stakeholder groups were invited to actively participate in discussions and not simply just observe proceedings. He also points to the decision by a large group of governments to walk away from negotiations of International Telecommunication Regulations, with the formal explanation given by the US for their decision not to sign the treaty referencing the inconsistency ‘with a multi-stakeholder model of internet governance’ and stating that ‘provisions on internet governance’ should be included in the redefined scope of telecommunication regulation.
Epstein argues that the bottom-up involvement of non-state actors in shaping agendas could not have evolved separately from the establishment of the IGF and the practices that were institutionalised through the IGF. He argues that “the IGF has an important symbolic meaning as an institutional framework that allows coexistence of competing political interests and values.”
The International Institute for Sustainable Development (‘IISD’) produced a review of the mandate of the IGF, where they stated that the IGF has developed a unique international policy dialogue function, “overcoming traditional reluctance to have public or private sector policy decisions openly discussed and debated at the international level”. With respect to the IISD’s own goal of promoting effective internet policy in pursuit of sustainable development, they say that “the IGF has been an extremely important arena for IISD’s work”. They list a number of indicators of progress where their work has influenced others through the mechanism of the IGF, claiming that “we would not have been able to advance our agenda without an arena like the IGF”.
The IISD report however does claim that they cannot be certain whether the IGF, and the discussions which take place within this forum, have had any impact on the quality of decision making in “more established internet governance forums”. This leads us to address some of the criticisms that the IGF has faced, and some of the drawbacks of its model.
The main critique of the IGF appears to be that it does not have any binding effect on participants; this makes it difficult to say with certainty that they have directly influenced any decisions. Milton Mueller argues that the creation of the IGF was “widely understood to be the kind of agreement that could get the WSIS out of its impasses; it allowed the critics to continue raising their issues in an official forum, but as a non-binding discussion arena, could not do much harm to those interested in preserving the status quo”. Perhaps more concerning is that many governments do not send representatives to meetings where decisions are not taken. This may be why some have argued that the IGF has not been successful in bringing issues to the attention of relevant bodies and the general public, and why they have had a lack of success in making appropriate recommendations. Indeed it is difficult to point to any concrete changes or decisions and say with confidence that the IGF has directly influenced these.
David Souter, who has been an active participant in the IGF, and attended many of their meetings, has stated that the IGF isn’t as wide-ranging in its perspectives as it ought to be. He argues that there is a broadly shared consensus about how the internet should be governed, and those who hold competing views tend not to go to the meetings. Similarly, he claims it is not as diverse as it would like to be. He points to the cost of travel and accommodation as reasons why a number of large countries are often poorly represented. This however may be mitigated somewhat in the future with a move towards a more virtual format.
How could it be improved?
Better engagement with decision makers
At the IGF meeting ‘Advancing the 10-Year Mandate of the IGF’, it was suggested they appoint a Special Advisor to the UN Secretary-General for Internet Governance, to act as the “political face” of the IGF, protecting the IGF Secretariat from external pressure, and ensuring it can concentrate on performing operational tasks. Crucially perhaps the report recognises that “governments react to seniority levels, and some have felt that having this type of appointee to represent the IGF would attract more high-level engagement.” Oftentimes those with the power to directly influence government policy are absent from IGF meetings. This is one of the reasons the IGF has struggled to influence governments and those on the international stage in general. Indeed some noticed that the 2020 iteration of the IGF saw a decrease in high-level political pressure, with only one Head of State in attendance. A similarly low turnout from the private sector at CEO level was also noted.
The UN Secretary General’s Roadmap for Digital Cooperation, published in June 2020, provides a number of suggestions to improve the IGF. One suggestion it makes is the establishment of a high-level segment and ministerial or parliamentarian track, helping to ensure more actionable outcomes. This would make participation in the IGF an opportunity to engage in discussion and information exchange regarding the preparation, adoption, and implementation of legislation on internet related public policy issues. The roadmap also suggests creating a strategic and empowered multi-stakeholder high-level body, building on the experience of the existing multi-stakeholder advisory group, which would address urgent issues, coordinate follow-up action on IGF discussions and relay proposed policy approaches and recommendations from the IGF to the appropriate normative and decision-making forums.
If there was also a way to more closely align with the UN General Assembly (‘UNGA’) for example, whilst still maintaining independence and keeping itself free from political pressures, the IGF may have the opportunity to influence UNGA resolutions. Not only would this help achieve the suggestion set out in the Roadmap of enhancing the visibility of the IGF and improving reporting to other UN entities, but these soft law instruments can be extremely influential at both the regional and international level. The European Court of Human Rights for example has in the past explicitly attached importance to soft law provisions in spite of their non-binding nature.
Some feel that there are currently high-entry barriers for newcomers. The suggestions to improve the IGF in this regard include: capacity-building at the national and regional levels; specific workshops/webinars; involvement of internet governance schools and programmes; and enhancing general communications and outreach capacities.
Implications for Artificial Intelligence governance
Artificial Intelligence governance today
As things currently stand there are no hard law regulations concerning Artificial Intelligence (‘AI’). Many have argued, and will continue to argue, that global legal measures are needed, especially when considering the ease at which AI can permeate borders.
There are already a vast number of organisations and national and international sets of AI ethics and governance principles worldwide, and this number is rising all the time. According to the OECD, there are also around 600 AI policy initiatives from 60 countries, territories, and the EU. The fragmented picture this paints of how things stand with AI governance today is exacerbated by the fact that there is already conflict amongst certain countries about the optimal approach to any potential regulation.
The dichotomy between the preferred US and European approaches illustrates this tension.
The US approach appears to be to endorse a largely laissez-faire approach to AI regulation, and they encourage others to follow this. In early 2020, the White House put forward a set of regulatory principles which endorsed avoiding the overregulation of AI technologies in the private sector. They stated that they hoped “Europe and our allies should avoid heavy-handed innovation-killing models and instead consider a similar regulatory approach”.
The EU however, as evidenced in their recent proposals for enacting harmonised rules of AI, seems intent on implementing a stronger regulatory framework, much in the same way they have approached issues of data protection, most notably through the GDPR. Some have cautioned that this is likely to be a lengthy, iterative, and highly complex undertaking.
Could an Artificial Intelligence Governance Forum be a positive step forward?
It is not just governments who are concerning themselves with questions of AI regulation. According to a PWC report, in recent years universities, NGOs and large companies have begun developing high-level frameworks and guidelines for AI governance. One issue is that there is seemingly very little co-ordination in how these issues are approached. An Artificial Intelligence Governance Forum (‘AIGF’) could perhaps help in this respect. Much in the way the IGF has done, it could create a multi-stakeholder space for open discussion and dialogue, allowing input from multiple sources, including government, NGOs, businesses and the public. This could help to encourage an aligning of regulation, or at the very least help encourage more of a consensus amongst how best to approach this difficult issue. It could also help make progress on some of the most contentious issues, for example definitions, which will be a necessary first step towards any governance regime.
The AIGF as a soft law mechanism
One of the difficulties with the traditional legal and regulatory approach focusing on legislation is the length of time it can take to respond to such a fast-changing area. A group of academics from Arizona State University have proposed that soft law mechanisms may be a promising alternative approach, at least in the interim. Soft law is defined as a program that sets substantive expectations, but isn’t directly enforceable by government. It has a number of advantages as a governance strategy for AI: it is not beholden to jurisdiction, it can be developed, amended and adopted by any entity, allowing it to be flexible and adapt to changing circumstances and technological developments.
This approach is not going to solve the issue, but it could be a valuable first step. If the end goal is a system of global legal regulation, an AIGF could help bridge the gap between discussion and action, bringing together different stakeholders and laying the foundation for a future move to legal regulation.
One of the major advantages any AIGF would have is that it can observe and learn from the experiences of the IGF. There is no doubt it would have its own difficulties and face critiques in the future, but perhaps it would not be setting off into the unknown quite in the way the IGF was when it was launched many years ago. Any suggested improvements to the current iteration of the IGF could arguably be examined with an eye to a future AIGF, to ensure it has the strongest possible starting structure from which it can evolve over time. To be truly effective as a soft law mechanism it would also likely have to have closer ties to, or greater influence within, the UN or other supranational bodies.
A recent KPMG report argues that given the pace of AI adoption globally, we could see a rush to regulate over the next 7-10 years and certainly there seems to be no doubt that we are moving towards AI regulation, likely on a global scale. An AIGF could provide a truly global space for discussion and collaboration between invested and affected parties, ultimately laying the foundation for successful future regulation.
Arizona State University, Soft-Law Governance of Artificial Intelligence, available at https://lsi.asulaw.org/softlaw/about/
Euractiv, Avoid Heavy AI regulation, White House tells EU, available at: https://www.euractiv.com/section/digital/news/avoid-heavy-ai-regulation-white-house-tells-eu/
Epstein, D., The Making of Institutions of Information Governance: The Case of the Internet Governance Forum, Journal of Information Technology (2013)
European Commission, Proposal for a Regulation of the European Parliament and of the Council Laying Down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act) and Amending Certain Union Legislative Acts, available at: https://ec.europa.eu/transparency/documents-register/detail?ref=COM(2021)206&lang=en
IGF Retreat Proceedings, Advancing the 10-Year Mandate of the Internet Governance Forum, available at: https://www.intgovforum.org/multilingual/index.php?q=filedepot_download/3367/711
KPMG, The shape of AI governance to come, available at: https://home.kpmg/xx/en/home/insights/2020/12/the-shape-of-ai-governance-to-come.html
Mueller, M., Networks and States: The global politics of internet governance, Cambridge, MA: MIT Press, 2010
OECD, National AI policies & strategies, available at: https://www.oecd.ai/dashboards
PWC, Comprehensive AI governance needed, available at: https://www.pwc.com/jp/en/knowledge/thoughtleadership/assets/pdf/comprehensive-ai-governance-needed-now.pdf
Review of the Mandate of the Internet Governance Forum: A response from the International Institute for Sustainable Development, available at https://www.iisd.org/system/files/publications/igf_mandate_review.pdf
Souter, D., Inside the Digital Society: Where next for the IGF?, available at: https://www.apc.org/en/blog/inside-digital-society-where-next-igf
United Nations Secretary General’s Roadmap for Digital Cooperation, available at: https://documents-dds-ny.un.org/doc/UNDOC/GEN/N20/102/51/PDF/N2010251.pdf?OpenElement
U.S. Intervention at the World Conference on International Telecommunications, Media Note, available at: https://2009-2017.state.gov/r/pa/prs/ps/2012/12/202037.htm
World Summit on the information society, Tunis Agenda for the Information Society, available at https://www.itu.int/net/wsis/docs2/tunis/off/6rev1.html
Blog posts from Trustees and associates of the One World Trust