25 October 2023
The European Commission launched a one-week consultation on 13 October 2023 to gather views on G7 guiding principles on generative AI. These principles will result in a code of conduct, which organisations could commit to. This is not a regulation and is not a legally-binding instrument. We are concerned that these principles may be insufficient when people are harmed by these AI systems as people will not have the possibility to complain to a regulator or go to a court.
ICCL responded to the consultation last week with the following response:
We thank the European Commission for the opportunity to provide feedback on the draft G7 guiding principles for Organizations Developing Advanced AI systems.
With regards to the draft guiding principles, we suggest two additional principles:
1. Do not develop or deploy advanced AI systems that violate human rights, undermine democratic values, are harmful to individuals or communities, facilitate terrorism, enable criminal misuse, or pose substantial risks to safety, security, and human rights.
2. Withdraw AI systems from the market if unable to mitigate risks, vulnerabilities and misuses, or unable to ensure the robustness, safety and security of systems throughout their lifecycle.
The first of these additional principles is part of the explanatory note. However, we emphasise that this should be included as an integral principle. The second principle is related to the first. An already deployed AI system that poses unacceptable harms which the organisation is unable to address should be withdrawn from the market.
We also suggest modification of one of the proposed guiding principles:
1. Principle 4 states that organisations should "work towards responsible information sharing and reporting of incidents among organizations developing advanced AI systems including with industry, governments, civil society, and academia." This principle should clarify that organizations should disclose near-misses incidents, not only serious incidents, so that other organizations can learn from them and prevent serious incidents.
In addition, we support the introduction of monitoring tools and mechanisms to hold organisations accountable to the guiding principles. Such a monitoring mechanism should be established at national and international level with a coordination mechanism established between these levels. Furthermore, the monitoring organisations should periodically publish transparency reports.