AI safety policy solutions and thought leadership

The summit underscored the deep divisions in how different stakeholders view AI regulation.

Does a rose by any other name smell as sweet? 

A comprehensive analysis of current AI agents reveals a significant lack of information about safety policies and evaluations.

Balancing ideological and geopolitical considerations with China competition

Meta's framework is a good start, but some glaring issues need improvement.

AI-themed advertisements are the first step toward a severe risk of “gradual disempowerment” for humanity.

The only way to make AI benchmarks even harder is to have them written and scored by AI.

AI reasoning systems are capable of formulating complex internal narratives that diverge significantly from their outward responses.

"The discussions at this tabletop exercise should be a wake-up call for government officials: the threat from AI is outpacing our preparedness."

The less people know about AI, the more likely they are to embrace it.

CAIP will be looking to hear 5 questions asked at Lutnick's confirmation hearing.

AI is a critical topic for any U.S. president to address.

If it's important enough to spend $500 billion on, it’s important enough to have a specific destination in mind.

CAIP supports whistleblower protections, safety planning, and shoring up cybersecurity

Broad Congressional support is needed for high-quality AI governance

CAIP’s response to the Department of Defense

A new paper on Best-of-N Jailbreaking demonstrates the inherent unreliability of reinforcement-based alignment

After two years of congressional deliberation on artificial intelligence, we need more than careful analysis

The continued lack of a legislative safety net for AI’s unintended consequences puts us all at risk

An op-ed in Tech Policy Press on AI's likely effects on musicians and how to guide them in a better direction.

CAIP looks forward to working with the Trump administration to promote common-sense AI policies

CAIP commends Congress for its hard work and timely achievements, but much remains to be done in FY2026

This legislation buys us a great deal of security at little or no cost to innovation

However, more work will be needed to develop appropriate laws

The United Healthcare tragedy reminds us that critical health coverage decisions cannot be safely delegated to inscrutable AI algorithms

We have hard evidence that AI is lying, scheming, and protecting itself, but developers don’t care

CAIP looks forward to working with David Sacks and the Trump administration

There’s more science to be done, but it’s not too early to start collecting reports from AI developers

CAIP's response to the AI Safety Institute

The most recent CAIP podcast explores four principles to address ever-evolving AI

The United States hosted the inaugural meeting of a growing global network

Twelve organizations call for action on bipartisan artificial intelligence legislation

Analyzing present and future military uses of AI

AI and the Future of Music: Innovation, IP Rights, and Workforce Adaptation

Slower AI progress would still move fast enough to radically disrupt American society, culture, and business

AISI conducted pre-deployment evaluations of Anthropic's Claude 3.5 Sonnet model

Op-ed in Tech Policy Press on AI safety testing and US-China competition

What AI music generation will mean for our creativity, current and future musicians, and the art in general

CAIP's response to the Department of Energy

Significant changes appear imminent in America’s AI policy landscape

For the US to lead on AI, AISI needs resources and formal authorization from Congress

CAIP's response to the National Telecommunications and Information Administration (NTIA)

Countless American sports fans are suffering emotional despair and financial ruin as a result of AI-powered online sports betting

CAIP conducted a survey of notable races this election cycle

Inspecting the claim that AI safety and US primacy are direct trade-offs

5 horror movie tropes that explain trends in AI

The FTC enacted its final rule banning fake AI reviews

Reclaim safety as the focus of international conversations on AI

The jury is out on the social and health effects of these tools

Unfortunately, not all AI agent applications are low-risk

Why is there still inaction in Congress?

Many voters prefer a careful, safety-focused approach to AI development over a strategy emphasizing speed and geopolitical competition

States cannot afford to wait any longer to begin setting AI safety rules

Creating a plan, anticipating challenges, and executing a coordinated response saves lives and protects communities

No one man, woman, or machine should have this much power over the future of AI

CAIP supports these reporting requirements and urges Congress to explicitly authorize them

The "godfather of AI" is concerned about the dangers of the technology he helped create

Policymakers and engineers should prioritize alignment innovation as AI rapidly develops

The rapid growth of AI creates areas of concern in the field of data privacy, particularly for healthcare data

CAIP was featured in Politico's coverage of the SB 1047 veto decision

It’s time for Congress to act

The Senate Select Committee on Intelligence bipartisanly, publicly, and loudly implored the American people and the private sector to remain vigilant against election interference

CAIP hopes that Walz and Vance will tell their fellow Americans where they stand on AI safety legislation

California Governor Gavin Newsom has vetoed SB 1047, a crucial bill to ensure the responsible development and deployment of AI

OpenAI's lobbying has expanded, and it's crowding out dialogue on safety.

An overview of AI explainability concepts and techniques, along with recommendations for reasonable policies to mitigate risk while maximizing the benefits of these powerful technologies

If Governor Newsom cares about AI safety, he'll sign SB 1047

The US is punching below its weight when it comes to funding its AI Safety Institute (AISI)

CAIP traveled to New York City to hear what local AI professionals have to say about AI's risks and rewards

Who is vouching for the safety of OpenAI’s most advanced AI system?

Engineers continue discovering techniques that boost AI performance after the main training phase

A new polls shows once again that the American public is profoundly skeptical of AI and worried about its risks

Jason Green-Lowe joined the Morning Rush TV show to discuss AI policy

Response to the initial public draft of NIST's guidelines on misuse risk

America’s best-beloved circles up a crew of technologists, humanists, and a law enforcer on what’s next for humanity in AI

CAIP calls on House leadership to promptly bring these bills to the floor for a vote

AI is spreading quickly in classrooms, offering numerous benefits but also risks

Out of 30 campaign websites reviewed, only 4 had even a single clear position on AI policy

Advancing Education in the AI Era: Promises, Pitfalls, and Policy Strategies

The voters have a right to know what their Presidential candidates will do to keep Americans safe in the age of AI

With the Senate returning today from its August recess, there are two strong bills that are ready for action and that would make AI safer if passed

The Center for AI Policy (CAIP) believes the 2024 Presidential candidates need to take a stand on AI safety

Last week, Brian Waldrip and I traveled to South Dakota, seeking to understand how artificial intelligence (AI) is perceived and approached in the Great Plains.

A 10-year-old girl accidentally hanged herself while trying to replicate a “Blackout Challenge” shown to her by TikTok’s video feed.

The Center for AI Policy (CAIP) organized and submitted the following letter to California Gavin Newsom urging him to sign SB 1047.

Yet another example why we need safe and trustworthy AI models.

Political campaigns should disclose when they use AI-generated content on radio and television.

The Center for AI Policy sponsored a mobile billboard to highlight the need for democratizing AI governance in the U.S.

Last Monday night (8/19/24), the Democratic Party approved its 2024 Party Platform. The platform’s general rhetoric hits all the key themes of AI safety.

Even if we plug the holes in our porous firewalls, there’s another problem we have to solve in order to win an AI arms race: alignment.

Reflections on a trip to DEFCON 2024

How will American AI firms respond to General Purpose AI requirements?

Human-quality speech presents a heightened risk that AI is used for fraud, misinformation, and manipulation

Why government oversight must complement corporate commitments

CAIP proposes that AI companies report their cybersecurity protocols against a set of key metrics

This step forward doesn’t mean that AI is free from sexual exploitation

Four bills advance to ensure c​ommonsense AI governance and innovation in the United States

Autonomous Weapons and Human Control: Shaping AI Policy for a Secure Future

Autonomous weapons are here, development is ramping up, and guardrails are needed

Meta essentially ran a closed-source safety check on an open-source AI system