Jersey Financial Services Commission

The first Global Dialogue on AI Governance should provide for standards on AI governance and ensuring basic reflections of diversity within it. It's usage and development should take into considerations of international standards of human rights and should be able to question it's users and creators in the event of certain abuses. The dialogue should also set standards and requirements for capacity building internationally to ensure that people are able to recognise and question AI creations, data sets etc. I would also like the Dialogue to explore the issues around accountability and whether AI or it's owners can have legal personality and what the consequences are for IP, or dangerous information that it produces.

• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight
• AI capacity-building

AI capacity building is critical in a world where deep fakes, fraud and criminality are possible through AI. In financial services there is a growing concern of the use of AI to evade lawful Client due diligence/ KYC methods for the purposes of fraud and money laundering/ financing of terrorism. Human oversight is key in ensuring the effectiveness of AI - so people using AI should know that it is a tool that they are responsible for. Social, economic, ethical, cultural, linguistic and technical implications of AI - when AI is being made to make decision that affect peoples social and economic lives, it must have clarity on the diversity, rights and nuances of different ethnicities, cultures, language and the technical capabilities of its users. AI is being developed by human beings who have their own biases and prejudices and these can inadvertently be integrated by AI so that when it is being used it will discriminate based on human biases and prejudices. As such if programmers, are not developing it from an inclusive lens we will continue to see marginalised groups further marginalised. This is especially important in health, e=education, financial services related tools. Protection and promotion of human rights, AI could be used to collate and derive insights for the promotion of human rights. In health it can be used to develop ideas around health and the recognition of discriminatory patterns in data, especially when prejudice in decision making. Transparency, accountability and human oversight are critical for the usage of AI, as there is required accountability. The use of AI in the context of crypto and digital assets requires full transparency, accountability and human oversight as the the systems are subject to ML/FT/PF. Further the data used to develop the AI must be transparent to ensure that the insights are accurate and the AI is not hallucinating.

Yes, in order for AI to be able to replicate human decision making or a tool to assist the decisions or actions it should be subject to the same standards the human beings are, which includes of recognition of social, economic, ethnicity, cultures and languages and varying technical capabilities. Capacity building is critical as not all people have equal levels of access to technology by virtue of background or disability. If it does not protect or promote human rights its usage for criminality can increase. As a regulator it is critical that there is AI that counters criminality or recognises the possible impact. Transparency, accountability and human oversight are critical aspects of recognising AI as a tool and the importance of having sound and representative data to mitigate the risk of hallucination.

There are infinite possibilities for evolving standards for CDD and eKYC for the mitigation of financial crime risk. The use of AI to recognise and exclude and prevent fraud Lack of capacity building to understand the potential efficacies. Having clear guidelines on standards when using AI Using AI that allow for active collaboration in investigations and recognition of financial crime especially when it comes to fintech and digital assets/ economy. Providing private and public sector AI products that recognise and report either SARs or other criminal events to the regulators from the private sector.

The dialogue is key to set international standards and understand where other countries are internationally. Further international cooperation is key in combatting financial crime.

Jersey AI forum FATF IOSCO IAIS UNWomen

Stakeholders can contribute on issues of data collection, data governance and how to create value in data. In identifying marginalised contributors who are a difficult group to acquire information from. Share their concerns and potential risks of AI so we can share ideas and build capacity on how to mitigate those risks.

People from the global south contribute in data but are not provided positions of governing AI and matters on data provision. AI organisations should not only take data from marginalised groups without including them in the purpose and governance of that AI model. For instance insights may be derived from data of women, or people who are low income but for the purpose of excluding them from financial services or housing. Young people are major users but are not considered in their opinion on the extent of the information that is there and the limitations of AI. It can be used for good but it might be a tool that is too powerful or can facilitate or encourage criminality amongst young people. People who use less common languages are less likely to be able to communicate with popular AI models. People who are living in poverty are likely to be excluded from useage and benefitting from AI models.

Discussion Panels Demonstrations of AI models Provision of opportunities to ask questions online or otherwise. Using case studies of current issues with AI and allowing a panel and attendants to resolve or analyse those case studies. A Hack-a-thon of designing solution based AI models (imagined or real).

AI Governance tools: Governing with Artificial Intelligence: The State of Play and Way Forward in Core Government Functions OECD