Quay lại Tin tức
Chính sáchAI Understanding tóm tắt

Oregon governor orders development of state-level AI safety regulations

Governor Tina Kotek has directed state officials to create safety guidelines for AI use, including third-party reviews and the potential implementation of a system-wide 'kill switch.'

4 min readRead the linked source
Source-provided image accompanying Oregon governor orders development of state-level AI safety regulations
Nguồn tham khảoNguồn đã ghi
Nhà xuất bản
oregonlive.com
Liên kết nguồn
oregonlive.comhttps://www.oregonlive.com/politics/2026/09/more-ai-regulation-ordered-by-oregon-governor.html
Loại nguồn
Nguồn được liên kết - trạng thái nguồn chính chưa được thiết lập.
Bối cảnhHiểu điều này trong 60 giây

Bắt đầu ở đây

Thuật ngữ chính

An toàn AI
Một lĩnh vực tập trung vào việc giảm các hành vi có hại, lỗi và rủi ro lạm dụng trong hệ thống AI.
Trí tuệ nhân tạo (AI)
Lĩnh vực rộng lớn của việc xây dựng các hệ thống thực hiện các nhiệm vụ yêu cầu nhận dạng mẫu, lý luận, ngôn ngữ hoặc ra quyết định.
Lan can
Các quy tắc, kiểm tra và kiểm soát nhằm hạn chế hành vi không an toàn hoặc không mong muốn của mô hình.
Tự kiểm traCâu đố về đạo đức AI

Chuyện gì đã xảy ra

Oregon Governor Tina Kotek issued an executive order on Wednesday mandating that state agencies develop a regulatory framework for the use of advanced artificial intelligence. The directive requires the state’s chief information officer to define 'frontier' AI models and establish criteria for third-party safety assessments before the state adopts such technologies. Additionally, the order tasks officials with evaluating the feasibility of a 'kill switch' mechanism to disable AI systems deemed unsafe.

Governor Tina Kotek’s executive order, announced on September 24, 2026, requires Oregon state agencies to establish a formal process for vetting AI tools. This includes the creation of qualifications for third-party reviewers tasked with assessing the safety of AI systems before they are integrated into state operations.

The order specifically targets 'frontier' AI models, the most advanced class of software. The state’s chief information officer is responsible for defining these models and determining which systems fall under the new oversight requirements.

A notable component of the directive is the requirement to assess the viability of a 'kill switch.' This mechanism is intended to provide an emergency shutdown capability for AI systems that are identified as posing safety risks. The governor’s office has acknowledged that state agencies currently utilize a range of AI tools, from standard chatbots to more complex systems used by technical staff.

Chi tiết nguồn: oregonlive.com ↗

Tại sao nó quan trọng

As federal AI regulation remains stalled, individual states are increasingly asserting authority over the technology's deployment within their jurisdictions. By mandating third-party safety reviews and exploring emergency shutdown capabilities, Oregon is joining a growing cohort of states attempting to mitigate risks associated with advanced AI. This move highlights the widening policy gap between state-level initiatives and the current federal stance, which has largely resisted new regulatory for the industry.

The order reflects a broader trend of state-level intervention in the absence of federal action. Governor Kotek explicitly cited the lack of federal oversight as a primary driver for the state's decision to act independently.

The initiative places Oregon alongside other states, such as California, that have recently moved to implement their own requirements. This fragmented regulatory landscape creates a complex environment for AI developers and state agencies, who must now navigate varying standards across different jurisdictions.

The focus on 'kill switches' and third-party audits underscores a shift toward more aggressive, precautionary governance models, prioritizing safety and control over rapid, unrestricted deployment of advanced AI systems.

Interactive Mechanism

Cơ chế tương tác: Nó thực sự hoạt động như thế nào

Khám phá công nghệ cơ bản đằng sau sự phát triển này một cách tương tác.

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:🛡️ Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
Kiểm tra khái niệm tương tác+10 Points
AI Ethics Quiz

Impossibility results in algorithmic fairness (e.g. Kleinberg et al., Chouldechova) show what?

Xem gì tiếp theo

The state’s chief information officer has 90 days to submit the proposed regulatory framework to the governor. Key unknowns include which specific AI technologies will be classified as 'frontier' models and whether the state will successfully identify a viable technical approach for a 'kill switch,' a concept that many technology experts have noted presents significant implementation challenges.

The 90-day timeline for the chief information officer to deliver the proposed regulations is a critical window. The resulting document will likely clarify the scope of the 'frontier' definition and the specific criteria for third-party reviewers.

The technical feasibility of a 'kill switch' remains a major point of contention. Observers will be watching to see if the state’s assessment concludes that such a mechanism is practical or if it remains a theoretical concept due to the inherent complexity of modern AI architectures.

The political context of the order, issued during a reelection campaign, may influence how the proposed regulations are received and implemented by state agencies in the coming months.

Hướng dẫn và câu hỏi liên quan

Đạo đức AITương lai của AIGiải thích về mô hình AIKiểm tra những gì bạn biết — thử một bài kiểm tra AI miễn phíTra cứu một thuật ngữ AI trong bảng thuật ngữ của chúng tôi
Tìm thấy điều này hữu ích?