Zeynep Tufekci

I strongly believe that third-party audits should be mandatory for major AI systems. These systems significantly impact our lives and society, and without proper oversight, they can perpetuate biases and cause unintended harm. Independent audits can help ensure transparency, accountability, and fairness, making AI technologies safer and more trustworthy for everyone.

As a technology scholar, I believe it's absolutely essential for frontier-model developers to publish their model cards before any commercial release. Transparency is key in understanding the implications and potential biases of these models. Model cards enable stakeholders, including developers, policymakers, and the public, to make informed decisions about their use. It's a critical step toward accountability and building trust in AI technologies.

I believe creating a global institute for AI safety is crucial. As AI systems become more pervasive and powerful, their potential impacts on society, privacy, and human rights demand coordinated oversight and regulation. Only a global approach can effectively address the complex and cross-border challenges AI presents.

I strongly believe that the liability for AI-caused harm should be shared between the developer and deployer. Both parties play critical roles in shaping how AI systems are designed, implemented, and managed. It is crucial to ensure accountability at every stage to prevent misuse and address any adverse consequences effectively.

Given the rapid advancement and potential risks of AI technologies, I believe it is crucial for a UN-led body to oversee compute-intensive AI development. Much like nuclear technology, AI holds significant potential but also poses great ethical and safety challenges. An international body could help establish guidelines and standards to ensure that AI is developed and applied in a manner that benefits humanity and prevents misuse.

I believe that establishing international AI ethics boards is crucial for guiding the responsible development and deployment of AI technologies. As AI becomes increasingly pervasive, it has the potential to impact society in profound ways, both positive and negative. An international board would help ensure that diverse perspectives and values are considered, fostering transparency, accountability, and fairness across borders. This collaborative approach is essential to address the complex ethical challenges AI presents, particularly in areas like privacy, bias, and autonomy.

As someone who studies the societal impact of technology, I believe it's crucial to establish international regulations to pause AI advancement. This pause will allow us to assess the potential risks, ethical considerations, and societal impacts of AI before it develops further. Without oversight, we might inadvertently create technologies that are harmful or uncontrollable.