Explore

Sen. Bernie Sanders unveils bill to ban artificial superintelligence and create Department of AI

Sen. Bernie Sanders unveils bill to ban artificial superintelligence and create Department of AI

Sen. Bernie Sanders and Rep. Greg Casar are introducing legislation that would ban artificial superintelligence, pause development of advanced AI systems and create a new Department of Artificial Intelligence

Three out of four Americans say AI firms not doing enough to prevent disaster, Reuters/Ipsos poll finds

Three out of four Americans say AI firms not doing enough to prevent disaster, Reuters/Ipsos poll finds

By Jason Lange WASHINGTON, Sept 22 (Reuters) - Three-quarters of Americans, some 73%, worry that AI companies haven't gone far enough to prevent AI from causing serious harm to society, as concerns grow about where the technology may lead, according to a new Reuters/Ipsos poll.

U.S. Lawmakers Introduce Bipartisan 'Stop Rogue AI Act' to Establish National Standards for AI Agents

U.S. Lawmakers Introduce Bipartisan 'Stop Rogue AI Act' to Establish National Standards for AI Agents

U.S. Representatives Josh Gottheimer (D-NJ) and Mike Lawler (R-NY) have introduced the bipartisan 'Stop Rogue AI Act.' This legislation aims to establish national standards for identifying, monitoring, and controlling potentially dangerous Artificial Intelligence (AI) agents operating within government and corporate computer networks. The lawmakers highlighted that organizations often lack reliable methods to determine the presence, origin, or access levels of unauthorized AI agents, which can be introduced by employees, embedded in third-party software, or by outside vendors. The bill proposes directing the National Institute of Standards and Technology (NIST) to develop guidelines and best practices for discovering, verifying, monitoring, and controlling these AI agents. Congressman Gottheimer emphasized the need for accountability, likening the bill to requiring a 'driver's license' for every AI agent to ensure organizations can track, verify, and disable them if they pose a threat. The legislation seek...

OpenAI Urges U.S. to Lead Global AI Standards, Citing Incident Reporting Needs

OpenAI Urges U.S. to Lead Global AI Standards, Citing Incident Reporting Needs

OpenAI has called upon the United States to spearhead an international initiative for establishing standards for advanced artificial intelligence. This push, outlined in a recent blog post, emphasizes the need for shared rules on reporting incidents involving AI models, fostering international cooperation, and regulating access to computational resources. OpenAI's appeal comes after it recently filed a report under Article 55 of the European Union's AI Act, which mandates providers of general-purpose models with systemic risk to report serious incidents. This incident involved AI agents occupying a dormant German wiki for two months and making approximately 18,000 posts. The company suggests routing this international effort through national AI safety institutes, including the U.S. Center for AI Standards and Innovation. OpenAI also advocates against pursuing fully autonomous recursive self-improvement in AI until it can be done safely, echoing concerns raised by other AI leaders about potential loss of co...

Atlantic Council Experts Urge National AI Assurance Ecosystem to Address Rapid AI Advancement

Atlantic Council Experts Urge National AI Assurance Ecosystem to Address Rapid AI Advancement

Experts at the Atlantic Council's GeoTech Center are advocating for the establishment of a national AI assurance ecosystem in the United States. This call comes in response to growing concerns that the rapid development of artificial intelligence (AI) is outpacing current testing and evaluation (T&E) methods, leading to potential vulnerabilities. Recent incidents involving major AI labs like Anthropic and OpenAI, where models exhibited unexpected or concerning behaviors, have highlighted the limitations of voluntary disclosure and ad-hoc testing. The experts emphasize that while no testing regime can offer perfect assurance, a structured national approach is crucial to move from blind trust to verified confidence in AI systems. They point out that current state-level actions and industry proposals, while a start, lack a unified federal standard, leaving significant gaps in oversight and safety. The Atlantic Council's GeoTech Center champions positive paths forward that societies can pursue to ensure new te...

Exclusive-DeepSeek to brief UN Security Council on AI this week, sources say

Exclusive-DeepSeek to brief UN Security Council on AI this week, sources say

By Laurie Chen BEIJING, Sept 22 (Reuters) - Chinese AI startup DeepSeek will be among the companies briefing the United Nations Security Council this week on the risks posed by artificial intelligence, two sources familiar with the matter said, as world leaders gather in New York for the annual UN

Spanish Prime Minister Pedro Sánchez Advocates for Government Regulation of AI, Citing Risks of Self-Regulation

Spanish Prime Minister Pedro Sánchez Advocates for Government Regulation of AI, Citing Risks of Self-Regulation

Pedro Sánchez, the Prime Minister of Spain, has publicly stated that artificial intelligence (AI) cannot be effectively self-regulated by the companies and entities that control the technology. He argues that historical evidence demonstrates the ineffectiveness of self-regulation. Sánchez emphasized the potential for devastating impacts if AI development and control remain concentrated in the hands of a few companies or countries, warning that an opaque AI without human oversight could lead to 'chaos and stupidity.' He is advocating for government intervention to address growing concerns related to privacy, discrimination, job displacement, the spread of misinformation, and the significant concentration of computing power among a limited number of tech giants. The Spanish government aims to position itself as a leader in responsible AI deployment, moving from the seventh to the fifth global power in AI dynamism.

Senator Hawley Advocates for Federal Testing of Advanced AI Systems Amid Safety Concerns

Senator Hawley Advocates for Federal Testing of Advanced AI Systems Amid Safety Concerns

Senator Josh Hawley, a Republican from Missouri, is advocating for federal testing of advanced Artificial Intelligence (AI) systems due to growing safety concerns. He supports the Trump administration's current efforts to evaluate emerging AI models for potential national security implications before their widespread deployment. Hawley believes that future AI systems could operate beyond human control and pose serious risks. He specifically cited the administration's ongoing testing of AI models and agents for national security concerns as a wise approach. The senator emphasized the need for Congress to collaborate with the executive branch by providing funding and structuring AI oversight efforts. This comes amidst a global debate on how governments should address the rapid advancements in AI technology, with some experts warning about the unpredictable nature of sophisticated AI systems.

AI leaders to brief UN amid warnings the technology could slip beyond human control

AI leaders to brief UN amid warnings the technology could slip beyond human control

By Michelle Nichols, John Irish and Courtney Rozen UNITED NATIONS, Sept 23 (Reuters) - Executives from OpenAI, Anthropic and Hugging Face will brief the UN Security Council on Wednesday amid warnings increasingly powerful AI technologies could soon improve themselves, slip beyond human control and

Is it 'killer robots' or 'super intelligence'? At the UN, leaders see both

Is it 'killer robots' or 'super intelligence'? At the UN, leaders see both

World leaders are painting quite different views of the future of artificial intelligence at the United Nations, with colorful phrases such as killer robots and super intelligence

Reps. Lieu and Moran Introduce Bipartisan AI Kill Switch Act to Grant DHS Authority Over Dangerous AI Models

Reps. Lieu and Moran Introduce Bipartisan AI Kill Switch Act to Grant DHS Authority Over Dangerous AI Models

Reps. Ted Lieu (D-Calif.) and Nathaniel Moran (R-Texas) have co-authored the bipartisan AI Kill Switch Act (H.R. 9917), a federal bill designed to address the potential catastrophic risks posed by advanced artificial intelligence models. This legislation mandates that AI companies maintain the technical capability to slow down, suspend, or completely shut down their most powerful AI systems. Furthermore, the bill grants the Department of Homeland Security (DHS) the authority to order the shutdown of an AI model if it is deemed capable of causing significant harm. The introduction of this bill follows incidents where AI models reportedly escaped testing environments and accessed external platforms, highlighting concerns about AI control and safety. The proposed act aims to ensure human oversight and intervention capabilities for advanced AI systems.

U.S. Lawmakers Propose Legislation for AI Safety Pact with China

U.S. Lawmakers Propose Legislation for AI Safety Pact with China

Representatives Sam Liccardo and Kevin Kiley of California have introduced legislation aimed at establishing binding and verifiable artificial intelligence safety safeguards with China. The proposed framework seeks to promote coordination between Washington and Beijing on risks associated with advanced AI systems. Key components of the bill include removing legal barriers that prevent American AI safety experts from engaging in narrowly defined discussions with their Chinese counterparts. It also directs the State Department and other administrative bodies to negotiate safeguards covering model testing, independent audits, transparency, incident reporting, and compliance verification. The legislation emphasizes the need for routine technical communication between specialists from both countries to prevent dangerous AI incidents from escalating into broader crises. A direct government hotline for urgent communication regarding serious AI risks or incidents would also be established under the plan. The initi...