White House Will Ad Hoc Decide Who Can Individually Access GPT-5.6

- The article reports that the White House will handle approvals for access to GPT-5.6 (a hypothetical future AI model) on a case-by-case basis, rather than setting universal rules. - This marks a shift from earlier approaches that tried to establish broad regulations or safety frameworks for advanced AI, toward ad hoc government control over who gets to use the most powerful models. - GPT-5.6 is not a real product name; it's used as a placeholder for the next generation of frontier AI models beyond GPT-4, likely referring to systems from companies like OpenAI, Google DeepMind, or Anthropic. - The background context is the ongoing policy debate in Washington about whether and how to regulate "frontier AI" — extremely capable models that could pose risks of misuse (e.g., bioweapons, cyberattacks) or systemic societal harm. - Key actors: the White House Office of Science and Technology Policy (OSTP), AI companies, and national security agencies. The article suggests this ad hoc approach gives the executive branch enormous discretionary power over AI access.