RECOMMENDATION FOR THE DEVELOPMENT OF AN OVERSIGHT BODY FOR DEVELOPMENT OF AI APPLICATIONS IN CANADIAN ORGANIZATIONS

18 mars 2018

Par John McNally, MA Candidate, Carleton University

Abstract:

 

AI development should be required to meet the same degree of external scrutiny as cyber-security mandates, with individual firms being forced to meet external standards. Additionally, the creation of an oversight agency within ISED or an appropriate government body would allow for the Canadian government to engage with firms, develop standards to ensure data protections/ethical issues are engaged with, and would allow government involvement in the cutting-edge research space of AI.

 

Values underpinning recommendation: Principles of responsible development, concerns surrounding data privacy, and desire to develop a standards set that lays the groundwork for healthy development of AI technologies in the interest of all.

 

Recommendation: The development of a national, or international, oversight body that conducts research, sets standards and develops a recognized regulatory framework that firms developing AI technologies in Canada must adhere to.

 

Arguments to justify recommendation:

 

  1. AI Policy development is a new field, and Canada faces an opportunity to take the lead in the development of standards to follow 'Canadian' values of equality and open-ness in the interests of greater prosperity. The development and dissemination of AI technologies by groups who do not share Canadian values may risk the development of a landscape in which data privacy and biases are viewed as ancillary rather than barriers to fair development. This scenario can be avoided if Canada takes a leading role in defining the environment for new technology design.

  2. AI technologies have short-term and long-term implications for humanity, and positive or negative developments depend entirely on the structures surrounding design. Industry leaders and figures have long cried for a regulatory mechanism surrounding AI development, but the potential impacts on establishing an innovation-focused industry make this suggestion difficult to implement.Instead, coordination with industry figures should rely on the creation of standards that developers would be encouraged to follow. Given the industry's desire to maintain social capital and create technologies that will benefit society, adherence to these standards should be developed in tandem with stakeholders to ensure the standards developed are rigorous, relevant and in the best interest of all.

  3. Public acceptance of new technologies is currently maintained by a lack of technical understanding of technologies and their scale of deployment. That will swiftly change in the first instances of an autonomous weapon killing a civilian in battle, or in a corporate deployment of AI technologies directly negatively affecting consumers. In order to foster an environment of greater understanding and acceptance, Canada should prioritize proactively having the answers to the questions that will be asked when issues arise. Creating a standards set also enables an oversight body to recognize these issues before they arise and avoid them, making a safer and more recognizable world for all parties.

  4. Canada's leadership in the fields of research and development, including the creation of the Montreal Convention itself, makes it a respected and well-positioned voice from which to thrust this debate into the mainstream. The risk of catastrophic misinterpretation of new developments is enormous, and reducing the risk of AI-opposition to be used as a political tool is crucial. Setting standards for oversight will allow Canada to set the standard for the world. Canadian leadership in research will directly translate into these values embedded in commercialized technologies, which will make the world a safer place and one more reflective of the ideals we strive for. The development of Canadian standards is a direct projection of Canadian soft power - one equivalent to the deployment of peacekeepers and acceptance of refugees, and therefore should be sought as a both an appropriate technological, ethical and political mechanism to make the world a safer place.

© 2017 Déclaration de Montréal pour le développement responsable de l'IA

Université de Montréal

Suivez-nous!

  • Grey Twitter Icon
  • Grey Flickr Icon