Session
Economic and social inclusion and sustainable development: What is the relationship between digital policy and development and the established international frameworks for social and economic inclusion set out in the Sustainable Development Goals and the Universal Declaration of Human Rights, and in treaties such as the International Covenant on Economic, Social and Cultural Rights, the Conventions on the Elimination of Discrimination against Women, on the Rights of the Child, and on the Rights of Persons with Disabilities? How do policy makers and other stakeholders effectively connect these global instruments and interpretations to national contexts?
Digital policy and human rights frameworks: What is the relationship between digital policy and development and the established international frameworks for civil and political rights as set out in the Universal Declaration on Human Rights and the International Covenant on Civil and Political Rights and further interpretation of these in the online context provided by various resolutions of the Human Rights Council? How do policy makers and other stakeholders effectively connect these global instruments and interpretations to national contexts? What is the role of different local, national, regional and international stakeholders in achieving digital inclusion that meets the requirements of users in all communities?
Round Table - Circle - 60 Min
Description: Judicial systems worldwide are using Artificial Intelligence (AI) to analyze large amounts of legal data to help lawyers identify precedents in case law, enable administrations in streamlining judicial processes, and support judges with predictions on issues including sentence duration and recidivism scores. The emergence of legal analytics and predictive justice has implications for human rights as AI systems’ opaqueness can go against the principles of open justice, due process and the rule of law. The IGF Open Forum will be a platform to discuss emerging issues at the intersection of AI and the Rule of Law. Participants including judges, lawyers, researchers, public prosecutors, representatives of international organisations and youth be invited to share the challenges and the good practices adopted in their jurisdictions concerning the use of AI within judicial administrations and the legal implications of AI for society from a human rights perspective. Participation of representatives from all regions of the world will be ensured, notably based on the survey of judicial operators launched by UNESCO in 2020 that received responses 1200 respondents from 100 countries. The panel discussion will be followed by the launch of an online training for judicial operators on AI and the Rule of Law. In the form of a Massive Open Online Course, this training will stimulate a participative dialogue with judicial operators on AI-related innovations in the judicial system, and court rulings concerning artificial intelligence. It will facilitate knowledge exchange and experience sharing among judicial operators on artificial intelligence, and existing norms and standards in the field. The course will underline the implications of AI for human rights, highlighting existing case studies and best practices that translate ethical principles into practice in terms of the use of AI in justice systems and in cases involving AI impacting human rights. The discussions with stakeholders will be used to inform the second version of an online training on AI and the Rule of Law that will be developed by Cetic.br / NIC.br, UNESCO and its partners.
1) How will you design the session to ensure the best possible experience for the online and on-site participants? Only online session is foreseen. The roundtable format will facilitate open discussion between different stakeholders. 2) If the speakers and organizers will all be online, how will you ensure interactions between them and the participants (including with on-site participants)? Moderated discussion on 3-4 key questions will be followed by the launch of the online training for judicial operators. Participants will be expected to deliver short introductory remarks, followed by moderate Q&A session. 3) Are you planning to use complementary tools/platforms to increase participation and interaction during the session? The session will be streamed on Cetic.br/NIC.br and UNESCO social media feeds and any questions from those platforms will also be communicated to the moderator in real time.
Cetic.br/NIC.br and UNESCO
Prateek Sibal UNESCO; Vanessa Dreier, UNESCO; Alexandre Barbosa, Cetic.br/NIC.br; Ana Laura Martinez, Cetic.br/NIC.br
- Hon. Justice Edward Asante, President of the Economic Community of West African States (ECOWAS) Court of Justice, Africa
- Katherine Forrest, former United States District Judge of the United States District Court for the Southern District of New York, North America
- Isabela Ferrari, Federal Judge, Brazil, Latin America
- Benes Aldana, President, National Judicial College, North America
- Nicolas Miailhe, The Future Society, Athens Roundtable on AI and the Rule of Law
Prateek Sibal (UNESCO)
Vanessa Dreier (UNESCO)
Ana Laura Martinez (Cetic.br | NIC.br)
16. Peace, Justice and Strong Institutions
16.10
16.3
16.a
16.b
17.8
Targets: 16.3 Promote the rule of law at the national and international levels and ensure equal access to justice for all. Provision of training to judicial actors on the implications of AI for human rights strengthens rule of law at the national level. Provision of training on the use of AI by courts in the administration of justice has implications for promoting access to justice through potential time saving achieved in administration of justice due to the use of AI systems for streamlining administrative processes in a manner that conform with human rights.
16.A Strengthen relevant national institutions, including through international cooperation, for building capacity at all levels, in particular in developing countries, to prevent violence and combat terrorism and crime. The project strengthens human and institutional capacities with respect to AI and the rule of law. It further builds networks between judicial actors across the world, particularly in the framework of South-South cooperation, using the MOOC platform, conferences and topic specific webinars that facilitate interactions between judges from different countries.
16.B Promote and enforce non-discriminatory laws and policies for sustainable development The MOOC and trainings address the concerns around bias and discrimination in AI driven automated decision-making systems being used by judicial systems and law-enforcement agencies. 16.10 Ensure public access to information and protect fundamental freedoms, in accordance with national legislation and international agreements The MOOC and trainings work towards protecting fundamental freedoms by facilitating knowledge exchange and strengthening capacities of judicial operators concerning AI and the rule of law. 17.8 Fully operationalize the technology bank and science, technology, and innovation capacity-building mechanism for least developed countries by 2017 and enhance the use of enabling technology, in particular information and communications technology An indirect impact of the MOOC would be to positively enhance the quality of the legal environment in which ICTs are developed and used, which may engender greater innovation and adoption of such technologies.
Report
There is a consensus on the critical importance of capacity building for judicial operators regarding AI and the rule of law. It requires moving beyond basic understanding to expertise and inclusive training for using and mobilizing tools and algorithmic processing across cases, both in criminal and civilian justice. It is also about equipping judicial operators with the capacity to bridge the gap from principles to practice.
Algorithmic bias and related transparency issues were identified as core challenges associated with AI in justice systems. The need to resort to such systems was thoroughly acknowledged, given the benefits they bring, while recognizing some of their main challenges. A better understanding of AI systems and bias, as well as continued global dialogue were identified as key to ensure that the tools used are fair and offer guarantees.
Moving from education to training by equipping judges and judicial operators with a common understanding of AI the Rule of Law.
The three pillars of the Athens RoundTable (policies, competence, and standards) need to be developed and implemented. A set of policies comprising from how courts operate to how companies operate needs to be designed and implemented, in combination with trust and multi-stakeholder cooperation in order to create a combination of self, soft and hard regulatory mechanisms, along with the tools to implement them.
Artificial Intelligence and the Rule of Law in the Digital Ecosystem Internet Governance Forum 2021
On 7 December 2021, the Internet Governance Open Forum on Artificial Intelligence and the Rule of Law in the Digital Ecosystem
was presented by Nicolas Miailhe (founder & president of the Future Society), Isabela Ferrari (Federal Judge at the Brazilian Supreme Court of Justice) and Benes Aldana (President of the National Judicial College), with opening words from UNESCO and Cetic.br | NIC.br, and moderated by UNESCO.
Nicolas Miailhe, founder & president of the Future Society, and co-partner in UNESCO’s Massive Open Online Course (MOOC) on AI and the Rule of Law started by stressing the critical importance of capacity building, and the mission of the course on AI and the Rule of Law for duty bearers. With this course, the Future Society aim to close gaps in two dimensions. One, is moving from education to training by equipping judges and judicial operators with a collective understanding of what AI and the Rule of Law, AI for the Rule of law, and the Rule of Law for AI means. Once we understand, we must move from a basic understanding to a more expertised and inclusive training on using and mobilizing the tools and algorithmic processing across cases in situations in criminal and civilian justice. Secondly, it is also about equipping judicial operators with the capacity to bridge the gap from principles to practice. Per Nicolas Mialhe “It has to be the result of ‘smart cocktails’ of self, soft and hard regulatory mechanisms.” Prateek Sibal, considered how the idea to work with a community of lawyers bringing AI cases to court could also be a way to go from principles to practice.
The digital revolution has gradually brought a variety of technologies into daily human interactions, and more recently AI has begun to be displayed in daily human activities. Legal disputes surrounding AI are therefore inevitable and judges need the training and the capacity to understand its legal and ethical implications. To provide examples, Benes Aldana, president of the National Judicial College, introduced the use of AI in medicine which can be expected to give rise to a multitude of malpractice scenarios such as misdiagnosis, improper treatment, bad outcome, privacy concerns, emotional distress caused by faulty diagnose, and potential discrimination if AI would prioritize the allocation of limited medical resources for patient care. There are other areas; how do defamation laws apply to AI generated speech? What ground rules should be in place when we use AI in sentencing? Per Benes Aldana, “The challenge for the judiciary will be challenging our most fundamental commitment of fairness, due process and truth. The future path of AI is neither predetermined nor beyond our influence, and the judiciary plays a crucial role in guiding the conversation on AI.”
Deep diving into the use if AI in the judiciary and specific use cases of AI in the Brazilian justice system, Supreme Court Judge Isabella Ferrari introduced the VICTOR tool, currently used in the Brazilian Supreme Court. To understand the benefits and threats of this tool, we need to understand how it the tool works and Brazilian’s unique mitigation environment. As Brazil has a heavy case load pendency, the Brazilian justice system already started implementing electronic lawsuits in 2010 and added a new requirement of general repercussion for the Supreme Court, which VICTOR now helps to categorize in a quick manner. Besides, the tool has been used to decipher massive amounts of files and recognize and categorize the most important parts and decisions necessary for the lawsuit. According to Judge Isabella Ferrari, “we have a super AI tool that understands everything that is happening in all the Brazilian courts, all lawsuits from the beginning to its end, it is a superpower, and every superpower comes with big risks.” Considering the risks, as VICTOR is still in the testing phase, parties are not advised and unaware of its use so cannot appeal on it. In rebutting the common argument that VICTOR only suggests, Judge Ferrari links this to algorithmic bias.
With regards to the impact of AI on the Rule of Law and related challenges, Benes Aldana identified Algorithmic bias as a core challenge associated with AI in justice systems and concluded that “we need a better understanding of AI systems and bias to ensure that the tools we use in the court are fair.” (For a more specific analysis on this issues, please refer to the "Reflection on gender issues" section of this report.)
Another area we should educate judges on is autonomous vehicles, involving injuries and death, with a set of complex legal questions in tort and criminal law and issues of causation arising from faulty AI decision-making. It will challenge judges in term of identifying the responsible party – owner of the machine, manufacturer, software developers, and various other contributors. Judge Ferrari added her perspective on human rights challenges in the Supreme Court of Brazil by reiterating the crucial importance of online courts and the use of AI in the Brazilian justice system to combat everlasting case pendency in Brazil, but to do so in a safe way. “We shouldn’t stop progress especially in countries where it is needed but must ensure the right path and continue global dialogue to guide us to this path.”
In discussing the path forward, the NJC will expand the reach of current training on AI and the Rule of Law by continuing effort such as UNESCO and the Future Society’s MOOC on AI and the Rule of law, provide courses and webinars on different topics of AI, and considers the Athens Roundtable as an excellent first step to share critical information on AI and justice. The Athens RoundTable aims to achieve a form of a call to action to understand the difference between AI that can be trusted and AI that cannot to advance justice. To answer this question, Judge Isabela Ferrari displayed the importance of understanding the basic of AI operations. What can I ask developers regarding transparency, why are they using decision trees, and can they use more transparent tools? We must demand as much transparency and accountability as possible and need to highlight and understand the issue of bias; both data and design bias. Besides understanding the basic operations, “we need to ask ourselves where can we make mistakes, not criminal law, not with peoples freedom, so where is it a safe space to use AI? AI can and should be used in various ways throughout the Brazilian justice system, but “where can AI help us in a safe place and where do we leave the work to judges?”
In finding the line between AI that can be trusted and AI that cannot advance justice, Nicolas Miailhi focused on the three pillars of the Athens RoundTable: policies, competence, and standards. As the MOOC is doing, we must equip judges and judicial operators to properly discharge their duties in a way that creates trust, we need move into more objective measurements including bench marking standards. Like the MOOC is doing, we need to equip judges and judicial operators with the capacity, competence. We need a set of policies - from how courts operate up to how companies operate. “Without trust and multi-stakeholder cooperation we will not be able to create the necessary cocktail of ‘self, soft and hard regulatory mechanisms along with the tools to achieve it’”.
When covering one of the Q&A questions during the session Judge Ferrari considered the intersection of AI and Intellectual property by considering the notions of creation, inventions and how this human ability is shifting toward technology. AI is already painting, creating music, and even creating perfume. Per Ferrari; “If an algorithm has created a perfume, who is the owner of the creation, is it the algorithm, the programmer?” By adding another field of law to the discussion, Judge Ferrari wants us to understand the complex scope of AI and how it will affect most fields of law. For example; civil liability in the case of autonomous cars and responsibility. There are thousands of questions without answer. “If in law we are used to look for an answer of a problem that exists, now with AI, we have no answers. This is the beauty and the beast of AI, and we need to discuss everything, in all the different fields of law, but it will take time.”
Answering to another question on prioritizing between regulatory versus standardization, Nicolas Miailhe agrees that both take time to be well-developed. Taking the lengthy timeline of the GDPR as an example to show that legislation takes time to materialize and often falls behind technology. Nonetheless, standards also take time. In addition to the issues of industrial requirements, political and economic reality has to be accounted for. That is why the Athens Roundtable tries to create bridges to help decide on the meaning of interoperable standards, converging international conventions on AI and human rights, self-determination, and pluralism. These questions are not easily answered, and we need to advance in parallel. Looking at the EU AI Act, the EU is trying to do exactly that by trying to advance a hard law instrument and having delegated the requirement to develop a set of standards to render this legislation smatter in avoiding overspecification. Prateek Sibal added that “soft regulation and standers are especially helpful because technology is evolving, often faster than the law, pinning with standards so you can update the stand. As we can see, the dialogue was built on each other´s views without any clear point of disagreement.