Page 73 - AI Standards for Global Impact: From Governance to Action

AI Standards for Global Impact: From Governance to Action

Table 2: Applying PDR framework to build trust in multimedia authenticity (continued)

Policy
Approach Method Benefit and/or outcome
Requirements
Response Enforcement Regulatory Penalties can be applied and rules Part 2: Thematic AI
interventions enforced through governments
enacting laws and regulations that
specifically address the techniques
and approaches that should be used.
They also address what happens
when such techniques are breached.
Explainability Use of explain- Decisions made by AI systems can
er-type algorithms, be checked to maintain a high level
AI model of reliability and trustworthiness. This
verification helps mitigate risks of IPR breaches.
methods and
information about
training datasets
used.
Content Clear and well communicated mech-
contestability anisms benefit individuals, helping
them dispute claims.
Dispute Platform bans Policing of problematic areas can be
mechanisms more effective and beneficial when
access to platforms and websites
that frequently host infringing
content is regularly removed.

Finally, the paper includes a set of practical checklists for use by policymakers, regulators,
and technology providers. These can be used when designing regulations or enforcement
frameworks, developing technological solutions, or preparing crisis response strategies. The
checklists are intended to help align stakeholder expectations, identify critical gaps, support
responsible innovation, and enable conformity with emerging standards and best practices.

The following recommendations are intended for policymakers, regulators, the media and
technology sectors, and standards bodies. Each can be operationalized swiftly to strengthen
multimedia content authenticity and build global trust.

The paper makes the following recommendations.

For policymakers and regulators:

• Consider the checklists provided in section four of the Policy Paper.
• Participate in and collaborate on standard development and alignment initiatives,
especially through multilateral forums to help promote regulatory alignment.
• Consider international standards when developing and implementing regulatory
sandboxes to test new technologies, policy approaches, and compliance models in
controlled environments.
• Adopt a PDR framework based on internationally recognized standards to structure
responses to content authenticity challenges.
• Consider data privacy and bias regulations to ensure AI-generated content respects user
rights and avoids discriminatory outcomes.

61