Center for Data Innovation

The Center argues that mandatory government-directed “kill switch” approaches risk creating cybersecurity, reliability, and national security risks and should not be viewed as a sufficient substitute for broader controls needed to manage autonomous AI systems.

The Center argues against calls to slow frontier AI based on ungrounded probabilities of catastrophic outcomes, emphasizing evidence-based safety research, practical safeguard approaches, and the need to clarify what is known versus merely feared.

The Center discusses the practical and legal issues involved in shutting down AI systems, including confidentiality protections and uncertainty around government authority, in response to recent events involving model access controls and negotiations.

The Center argues that Congress should address specific shortcomings in the Deterring American AI Model Theft Act of 2026 approach so it deters harmful extraction without discouraging legitimate research or slowing U.S. innovation.

The Center argues that policymakers should distinguish between different types of AI-related copyright/trademark claims and treat the CNN-Perplexity lawsuit as legally distinct from training-data disputes; it contends existing legal tools may already address the underlying issues.

The Center presents a report asserting that many public concerns about AI data centers stem from measurement and management frameworks rather than AI infrastructure scale itself, and it outlines policy directions focused on systemic impacts such as electricity, grid access, pricing, reliability, and water.

The Center discusses how different approaches to using publicly available information (including personal data) shape the future trajectory of AI development and regulation across jurisdictions.

The Center summarizes a Google-sponsored Washington, D.C. pre-summit convening with The Dialogue ahead of the India AI Impact Summit 2026, focusing on adoption, compute, and governance as cross-cutting themes for U.S. strategic engagement.