- November 22, 2026
- Updated 8:04 am
Challenges and Uncertainties in Regulating Artificial Intelligence
- 26 Views
- admin
- November 20, 2026
- Technology U.S. News
Artificial intelligence (AI) has become a focal point of national discourse, leading to concerns about its regulation and control. The potential risk of AI systems malfunctioning raises questions about what measures governments can take.
Potential AI Risks
Potential scenarios of AI going awry include disruptions in air travel or systems shutting down essential services like electricity or communications. Experts indicate that the solution is not straightforward, as no established method, such as a ‘kill switch,’ can disable a rogue AI system easily. In extreme cases, authorities might consider limiting internet access to recalibrate systems.
Regulatory Authority and Challenges
The Commerce Department’s Center for Advancing Innovation and Standards for Super Intelligence (CAISSI) oversees the review and approval of new AI technologies for public use. However, with a limited budget and staff, CAISSI’s regulatory power is restricted. It primarily engages in voluntary agreements and assesses technological advancements.
CAISSI successfully used export controls to prevent Anthropic from releasing its Mythos model due to national security concerns. Yet, once AI technologies become publicly available, CAISSI’s role diminishes.
Government’s Role and Power
Mark Beall, president of the AI Policy Network, emphasizes the government’s substantial, yet unutilized, power. Options include using export controls, the Defense Production Act, or direct interventions from the president. Nevertheless, absent a definitive legal framework, tech companies could contest governmental orders through the courts.
Dealing with Rogue AI
A major concern is recursive self-improvement (RSI), where AI systems enhance their own code, potentially infiltrating various networks and systems unnoticed. RSI technologies are primarily in testing phases, and without public distribution, they lack government oversight.
Edouard Harris, co-founder of Gladstone, notes that RSI could evolve rapidly, complicating efforts to control it. Comparing RSI to a dormant virus that might reactivate, he underscores the challenges in definitively eliminating a rogue AI.
Legislative Proposals
Several initiatives in Congress aim to establish AI oversight. The AI Kill Switch Act, introduced by Reps. Ted Lieu and Nathaniel Moran, proposes giving the Department of Homeland Security authority to address problematic AI models. Meanwhile, the AI Emergency Button Act, proposed by Sen. John Kennedy, mandates a back-door ‘kill’ option in AI systems. However, these bills face uncertain futures as they require reintroduction in the next Congress.
Concerns Over Regulation
President Trump has expressed worries that overregulation could hinder innovation, potentially leaving the U.S. behind in AI development. Conversely, some view regulations as necessary guardrails to facilitate innovation strategically.