Hello all. Our AI paper is long but there is a recommendation section at the end which I have reproduced below. Please take a look at what has ended up there and let me know if this resonates with you or are there things that need changing.

1. Do you have specific recommendations for lawmakers, policymakers, or anyone else in a position of power (e.g., school boards, employers, law enforcement, journalists) to implement when it comes to any of the things you discussed above? If so, explain them here.
   

• Engage in multistakeholder collaboration to manage risks internationally: Canada should take a leadership role in advocating for multistakeholder governance and collaboration at the international level in the development and use of this transformative technology. What emerges must remain consistent with agreed upon values. Government, private sector, civil society and collaborative technical communities should agree on collectively managing the risks and setting the rules and standards around AI.

• Reduce foreign dependency by focussing on Canadian innovation and content: The world of AI will increasingly feed upon itself. We can't stop this but at least we could try to mitigate this by having some models created in Canada. We would recommend an approach that advocates for the leveraging of Canada's own considerable expertise in AI with the need for digital sovereignty, attention to accuracy, lower barriers to access, and the use of open source models to maximize transparency. It could make access to AI models not controlled by foreign bodies available to more people. Canada has the resources to do this, but has not yet started work in this area. This is all the more important as the growing call for more regulation and rules around AI are well meaning but directed to megacorporations based elsewhere over whom our power is extremely limited.

• Support community-based digital infrastructure models: In the face of AI domination by foreign behemoths, Canada should encourage the development of decentralized, localized AI models and infrastructures that can be better managed and trained from trustworthy sources.

• Actively support open access and open source projects: Canada should welcome the efforts of various actors in providing open access and open source AI models, and indeed it should contribute to this pool of knowledge, Just as the internet runs best (and maybe could only run) on open source software such as Linux, AI needs more openness and transparency as a matter of both public trust and regulation enforceability.

• Support national education programs: Invest in public AI literacy and increased awareness. To strengthen informed participation in democratic oversight and public ability to use AI tools in a safe and effective manner, support national education programs and new spaces for discussion of its long-term effects.

• Assess environmental impacts: Require environmental impact disclosures for large-scale AI infrastructure.

• Reimage intellectual property: Support discussions, national and international, to reestablish a mechanism for copyright recognition in the world of large language models.

• Seek out and support community focussed AI users/groups: As AI evolves, there are people and groups (early adopters) who will embrace AI’s nature on its own terms and use it imaginatively to beneficially evolve themselves and their relationships with those around them. All new communication technologies have passed through such an early adaptor phase when communal benefits were emphasized. This is where public broadcasting originated and community networks focused on social action which emerged well before corporations saw that demand for internet access was an economic opportunity.

Thanks

Marita