AI Watch

AI tool poisoning exposes a major flaw in enterprise agent security– venturebeat.com

pgnewser May 11, 2026 0

News Source
EXCERPT:

AI agents choose tools from shared registries by matching natural-language descriptions. But no human is verifying whether those descriptions are true.

I discovered this gap when I filed Issue #141 in the CoSAI secure-ai-tooling repository. I assumed it would be treated as a single risk entry. The repository maintainer saw it differently and split my submission into two separate issues: One covering selection-time threats (tool impersonation, metadata manipulation); the other covering execution-time threats (behavioral drift, runtime contract violation).

That confirmed tool registry poisoning is not one vulnerability. It represents multiple vulnerabilities at every stage of the tool’s life cycle.

There’s an immediate tendency to apply the defenses we already have. Over the past 10 years, we’ve built software supply chain controls, including code signing, software bill of materials (SBOMs), supply-chain levels for software Artifacts (SLSA) provenance, and Sigstore. Applying these defense-in-depth techniques to agent tool registries is the next logical step. That instinct is right in spirit, but insufficient in practice.

SpaceX wants to launch a million satellites. Here’s how that could impact the atmosphere and the night sky– www.cbc.ca

pgnewser May 11, 2026 0

News Source
EXCERPT:

Most of humanity has a great propensity to think in the short-term, but generally, long-term considerations — air pollution, deforestation and emissions, for example — just aren’t our thing.

That’s partly why scientists are deeply concerned about a recent SpaceX proposal to launch one million satellites — data centres — into orbit around Earth.

Their concerns range from losing the natural night sky, to losing access to space, to the environmental impact on our atmosphere.

At the moment, there are roughly 16,000 satellites orbiting Earth, 14,000 of which are active. SpaceX is responsible for more than 8,000 of them.

Virginia School District Injects DEI Into AI Tools– thefederalist.com

pgnewser May 11, 2026 0

News Source
EXCERPT:

A school district in Virginia is using a diversity, equity, and inclusion (DEI) filter on its use of generative AI (GenAI) in schools to inject a race-focused lens into teacher and administrative decision-making.

According to documents obtained by Defending Education, Albemarle County Public Schools (ACPS), a far-left school district surrounding Charlottesville, and which includes Thomas Jefferson’s Monticello, has built its GenAI policy based around its “anti-racism” policy, which is inherently discriminatory against white students in favor of other races.

“While schools should engage in good judgement and do their due diligence when it comes to Artificial Intelligence integration, the fact that the district is vetting AI based on its compliance with diversity, equity, and inclusion should be concerning for parents,” Rhyen Staley, Director of Research at Defending Education, told The Federalist. “By only allowing the use of AI and information sources that reflect a leftwing political bias, district administrators are setting a precedent that is harmful to the learning process and neutrality of schools.”

Fears of an AI breakthrough force the U.S. and China to talk– www.latimes.com

pgnewser May 11, 2026 0

News Source
EXCERPT:

CHONGQING, China — Three years ago, in the idyllic town of Woodside south of San Francisco, the United States and China held their first high-level talks on the dangers posed by artificial intelligence. President Xi Jinping and his longtime foreign minister appeared serious in their conviction that a channel should be a established between Beijing and Washington — a red phone for AI in case of emergencies.

They authorized a diplomatic effort that would begin in 2024 in Switzerland, only months before the U.S. presidential election. A large U.S. delegation arrived with high hopes that were abruptly dashed, according to four sources who attended the talks. The Chinese contingent dismissed American concerns over runaway AI as academic, almost theoretical, quickly turning the conversation to export controls seen in Beijing as yet another U.S. effort to hold China back.

“They naturally view any American diplomatic initiative involving limitations or restrictions of one flavor or another on a capability as being a trap,” Jake Sullivan, U.S. national security advisor under President Biden, said in an interview.

Make America AI Ready: Strengths, Weaknesses, and Recommendations– blog.citp.princeton.edu

pgnewser May 11, 2026 0

News Source
EXCERPT:

The Trump administration has made artificial intelligence a centerpiece of its economic agenda, promising to retrain a workforce it says must be ready to compete in an AI-driven future. One early piece of that effort: a free text-message course from the Department of Labor (DOL) and private partner Arist called, “Make America AI-Ready”, is a useful start on the journey to AI literacy for all Americans. This seven-day long, 10-minute-per day course which frames itself as “your AI 101” is accessible, technically informative, and engaging (see below for the full contents). Here we analyze its strengths, lay out a few weaknesses we think should be addressed in the current version, and elaborate some stretch goals for an “AI 201” course that would build upon the original.

AI Models Near the Point Where They Can Build a Better Version of Themselves – PJ Media– pjmedia.com

pgnewser May 7, 2026 0

News Source
EXCERPT:

“My prediction is by the end of 2028, it’s more likely than not that we have an AI system where you would be able to say to it: ‘Make a better version of yourself.’ And it just goes off and does that completely autonomously,” Jack Clark, who heads The Anthropic Institute, told Axios.

Clark, co-founder of Anthropic, says his institute is seeing signs of “AI contributing to speeding up the research and development of AI itself,” a process known as recursive self-improvement. 

Clark adds, “It’s always been the case that humans outside the technology need to come up with the ideas that they then put back into it. What happens if we have a technology that can generate ideas within itself for how to improve itself? That’s a new concept.”

Too fast, too soon. The speed with which AI systems are evolving is far outstripping our ability to gauge the impact on humans and society. Lots of good things can happen in medicine, biology, and other sciences where AI is already making a big impact. The speed and autonomy of artificial intelligence models promise an abundant future.

Anthropic AI says deal struck to use SpaceX data centres– www.cbc.ca

pgnewser May 7, 2026 0

News Source
EXCERPT:

The audio version of this article is generated by AI-based technology. Mispronunciations can occur. We are working with our partners to continually review and improve the results.

Anthropic on Wednesday said it had reached a deal to tap the computing resources of Elon Musk’s SpaceX, marking a détente with its one-time critic and ‌a boost for both companies in the high-stakes artificial intelligence race.

Under the agreement, Anthropic will use the full computing power of SpaceX’s Colossus 1 facility in Memphis, Tenn., which houses more than 220,000 Nvidia processors and will give the Claude chatbot maker 300 megawatts of new capacity within a month.

The deal gives the IPO-bound SpaceX a marquee customer as it looks to sell investors on its AI ambitions, while helping Anthropic ​ease capacity constraints following a surge in demand for products such as its AI coding ​tool, Claude Code.

The announcement came as Anthropic held a developer day in San Francisco on Wednesday, where it unveiled a new Claude AI feature called “dreaming,” meant to help its AI systems learn by reviewing work between sessions, spotting patterns and updating ​files that store user preferences and other context.

AI is splitting the housing market in two: Bay Area luxury homes up 13%, affordable ones collapsing– fortune.com

pgnewser May 7, 2026 0

News Source
EXCERPT:

As AI mints new millionaires, billionaires, and even trillionaires, it’s also threatening to replace entry-level workers and sparking fearful chatter of the “permanent underclass.” There’s no place that’s more evident than in the Bay Area, at the heart of Silicon Valley, where technology is wedging a deeper divide in the K-shaped economy, especially in the housing market.

A new Redfin report found that since the launch of ChatGPT’s first model in Nov. 2022, luxury home prices in the region—classified as those selling between $3.1 and $7.6 million—have jumped 13.4%. At the same time, home values for lower-end properties in the Bay Area—those $535,000 to $615,000—have fallen by 3.8%.

“Some owners of lower-end properties have missed out on the AI boom, with home prices in the most affordable Bay Area zip codes declining over the past two years,” Yingqi Xu, Redfin senior economist, said in a statement. “It’s another sign of the K-shaped economy taking shape in the Bay Area, with AI lifting the fortunes of some households and neighborhoods much more than others.”

Many Americans today are grappling with the sobering reality of high mortgage rates, inflated home prices, and a housing stock shortage. Many are delaying homebuying by a near decade from just a few years ago, as the median age of the first-time homebuyer hit 40 in 2025, up from just 33 in 2021.

Even brief AI use could hurt your ability to think, a new study finds– www.digitaltrends.com

pgnewser May 7, 2026 0

News Source
EXCERPT:

A new study from researchers at Carnegie Mellon, MIT, Oxford, and UCLA suggests that using an AI chatbot for just 10 minutes could negatively impact your ability to think and problem-solve. And honestly, the findings are a little alarming.

As reported by Wired, the researchers asked participants to solve problems, including simple fractions and reading comprehension tasks. Some participants were given access to an AI assistant that could solve the problem for them.

When the AI was suddenly removed, those participants were far more likely to give up or get the answer wrong. In other words, the moment the AI crutch was gone, people struggled.

AI and the Treasury Debt Market | American Enterprise Institute– www.aei.org

pgnewser May 7, 2026 0

News Source
EXCERPT:

The market for Treasury debt is challenging to interpret even when the environment is calm, much less when multiple geopolitical and technological disruptions are present simultaneously. Two papers released in the past year offer differing perspectives on a consequential unknown in the current market, which is the expected effect artificial intelligence (AI) will have on future productivity growth.

The first of the two relevant papers, co-authored by Isaiah Andrews and Maryam Farboodi, examined market signals in 2023 and 2024 around the releases of updated AI models from five leading developers (OpenAI, Anthropic, Google Deepmind, xAI, and DeepSeek). A main finding was that the nominal interest rates for long-term Treasury and corporate debt fell after these models became public by a statistically significant 12 basis points.

Leftist Censorship Groups Advise AI Company Anthropic– www.dailysignal.com

pgnewser May 6, 2026 0

News Source
EXCERPT:

Over the past decade, the Left has cultivated a censorship industrial complex of “experts” on “extremism” who try to bully Big Tech and corporate America into blacklisting conservatives over hot-button cultural issues such as LGBTQ+ orthodoxy and parental rights.

While the censorship industrial complex has suffered setbacks, it enjoys a persistent influence—notably at Anthropic, the major AI company behind the chatbot Claude.

Anthropic openly touts its relationship with four branches of the censorship industrial complex, and each of them has ties to the Southern Poverty Law Center.

The SPLC, the left-wing smear factory that pioneered the censorship strategy, now faces federal charges for allegedly lying to banks while attempting to conceal payments to members of the Ku Klux Klan, but Anthropic is not reconsidering its work with the SPLC’s allies.