Cat in the AI Hat

I think that the first Global Dialogue on AI Governance would be considered a success, if we realise that there is more than frameworks and laws required to govern and implement AI. In addition to that people need to be at the table who are not experts in the technical side of AI, Governance Risk and Compliance Experts, Lawyers, Solicitors and High Ranking High Achievers or experts with titles. AI Governance has to be by everyone, for everyone and including those who traditionally dont have a voice, cant afford to study or have titles or high ranks have just as valuable input to this discussion and should not be discounted. Creating an AI Governance solution for the world is not going to come about by creating new rules and frameworks on existing rules that AI has already changed before we get anywhere. Its about making sure that those using it, affected by it or otherwise have been changed by it have a voice here too. AI Governance cannot be done in isolation nor by frameworks and guidance that is out of date before it is even complete. A different approach is required and a successful outcome will recognise that a non traditional approach is required to help everyone and not just those who proclaim expertise

• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight
• Safe, secure and trustworthy AI

AI use in some places is endemic, people without AI literacy and guidance are using AI for uses that are not covered in any framework, from the proliferation of small and other business using AI for ad and content generation, through to uses in music and art without any thought for artistic people and their trades. We also have a situation where language and idioms and culture is being eroded by LLMS being trained on predominantly US based language, numeracy and culture. Non verbal language such as BSL or smaller languages such as Welsh, Cornish and Gaelic are being eroded. Americanisation of words is happening all over the world because it is default installation and in some cases cannot be changed either. We have seen studies that AI is already male focused and treats women differently. This also applies to LGBTQ and people of different races and religions. Decisions guided by AI are being taken without any consideration of these effects Usage of AI is that endemic that even the writing of Victim Impact Statements which are intensely personal are now being done with AI. The creation of a UN AI task force,for education purposes as well as for enforcing across the world the usage of AI is needed. Current AI legislation is outdated before it is enacted and a centralised taskforce that helps, practical usage, gives guidance, builds literacy, makes sure language rights, culture, art, religion and privacy are protected . Nothing like this exists currently. We are currently further increasing AI Colonialism and AI Imperialism, and this leads to a repeat of things already seen in the dark days of human history. This is our one chance to stop that and in order to do this we are going to need to think differently.

We have to make sure we dont go for frameworks, and siloed thinking. Discussion is stuck within a gate kept AI classicism that surrounds those who are known for STEM or other scientific pursuits. There arent humanitarians involved, low skilled or people from areas inc homelessness, poverty, victims of AI related incidents, non qualified or practical practitioners who cannot afford or who otherwise will not respond to this. There are no non traditional approaches that challenge algorithmic and AI life demanding more frameworks that are unwieldy and are out of date before they are implemented. In short emerging issues are that those that should be in the room are not, have no voice and cannot change what is going to affect them. I bring lived experience of a number of different areas not considered and we are trying to impose or create without that input in favour of creating something that may look great on paper, but needs to have a flexibility, budget and dedication we have never seen. AI model growth and life cycles are 3-6 months. We have nothing currently that can even work with that and needs to constantly evolve. We also genuinely need to consider that some old fashion methods such as On Prem Servers, not cloud connected, and even pen and paper based things maybe required to ensure that we do not become entirely dependent on AI for everything and in turn lose ourselves and our humanity, spending millions and billions in an attempt to keep up with the jones's that makes us all actually worse off. There are growing movements eschewing AI and that is something that needs to be catered for and considered within this conversation too.

In the UK, we have a government that is relying on over 20 different regulators and a framework that no one understands to implement AI safety. There is no desire from the government to have central singular evolving legislation and the frameworks mean nothing for anyone outside of AI spheres. People have to navigate complexities they dont understand and have to go through Court in order to be heard which immediately prices people out of trying to exercise and defend their rights. Progress here is slow and due to the gaps in not only legislation and enforcement and the slowness to react its very possible to create AI that will change the world, or even harm people, be widely adopted and nothing will ever be done about it until after the change has happened. The over reliance on US tech stacks and the absolutely stunning ignorance of EU legislation in the EU AI Act means that British Business is danger of falling apart massively, and when it does there is nothing or no one to pick up the pieces. The usage of AI has already permeated massive parts of life and there is no one monitoring or putting a stop to things, the proliferation of frameworks makes effective governance impossible and there is zero help for anyone affected by AI. No restitution funds, no holistic care, no support. There is no mass AI literacy or Privacy Programmes, just a blithe assumption that people will obey or remain in frameworks that are utterly useless, redundant, outdated and help no one, giving the illusion of safety when actually there isnt any, and no one to enforce it until way after the fact, damage already done.

Getting out of the need for GRC and Business protection frameworks, renaming it to AI Safety, as opposed to governance, investing and making the needed changes to cover everything and also as mentioned the need for mass education programmes and AI Literacy courses for everyone, not provided by AI companies or business, but standardised to be taught to the whole world

I dont think there are any currently as the problem hasnt been idenfied properly and as such there are not the organisations who can bring value to this discussion. We would need to create them as part of this process.

Mass Advertising, Social Media, shorter broken down aspects that do not cover everything all at once, and making sure that it is worded in a way that isnt too formal for greater response

LGBTQ People. Neurodivergent, those with non verbal language inc Sign Language, Those with digital barriers to participation. Working with charities of abuse, homelessness, multiple complex needs and others to reach communities that otherwise wouldnt be included would be useful to ensure their voices are heard.

Steering away from the technical, making it relatable, use slang, foster connections use everything from short format videos, that arent staged through to viral marketting, connect with the likes of Spud Bros or other vetted social media stars to promote strategy and discussions around this and consider working with others for other spaces.

There currently isnt anything that works. ISOs, EU AI ACT, NIST, OECD and the rest all favour business or after the fact 'governance' they are all outdated and come with supplimentary industries that try and sell other augmentations and implementations such as IAPPS AIGP that is only accessible if you have money and time to study for, meaning it becomes only useful for those with money. Happy to sit down and design with others something that does, but at the moment there isn't anything