Tag: ai

Sanders calls for artificial superintelligence ban after OpenAI CEO interview

Sen. Bernie Sanders (I-Vt.) on Monday called for permanently banning artificial superintelligence in response to recent remarks made by OpenAI CEO Sam Altman.

Altman told Politico’s “Decoded” podcast in an interview last week that society will face a “coming tidal wave of cybersecurity problems” coming from open-source AI models. He also said that people will “have to just accept some fairly severe cyber incidents from open models, in exchange for the liberty that comes with that.”

“No, Mr. Altman. We won’t accept a severe AI cyberattack in exchange for ‘liberty,’” Sanders wrote on the social platform X, later asking, “Mr. Altman: What do you mean by that?”

Sanders wrote that Americans “are not willing to accept disaster ‘in exchange for the liberty’ of making Big Tech Oligarchs even richer and more powerful,” adding that tech developers working on AI “are not above the law.”

“If you and other Big Tech Oligarchs make products that commit crimes or disasters, we will hold you accountable for breaking the law — no matter how much wealth and power you have,” Sanders continued. “We must PAUSE advanced AI and permanently BAN artificial superintelligence before it is too late.”

Sanders and Rep. Greg Casar (D-Texas) last month introduced the Ban Artificial Superintelligence Act, which would institute the permanent ban while temporarily pausing “advanced AI development” until federal regulators establish safety standards.

The bill would also direct the U.S. to seek diplomatic agreements to prevent the development of superintelligence and establish a Cabinet-level federal agency to focus on AI safety rules.

The bill is the latest of several proposals floated by progressives in Congress to slow down AI development amid recent cybersecurity incidents and hacks involving rogue AI software, including an unsuccessful hack of the Department of Education’s website and a hack of Australia’s health department.

The chorus of concern grew after former OpenAI and Anthropic researcher Jacob Coxon warned last month that AI could pose an existential threat to human life on Earth.

Sanders, who sided with Pope Leo XIV’s warnings about AI’s impact on the art world, slammed President Trump over his response to developing AI.

“Look, it is embarrassing, really, I think, going beyond politics, how ignorant this president is about the dangers of AI,” Sanders told MS NOW last month.

Trump said he would appoint an “AI czar” and has launched a “Super Intelligence Force” in response to the technology’s development, but he has praised AI’s possibilities despite the bipartisan calls for regulation.

The president has sought to rebrand AI as “super intelligence,” and he recently signed a pledge with tech executives that put the name in writing across the executive branch.

Keep reading

Flock CEO Garrett Langley to Critics of His Cameras: ‘We Shouldn’t Pause Progress’

Garrett Langley, founder and CEO of license plate camera maker Flock Safety, defended his company’s surveillance technology onstage at TechCrunch Disrupt 2026 in San Francisco this week, even as Flock contends with canceled contracts and layoffs nationwide.

TechCrunch reports that speaking during a panel interview, Langley pushed back on critics who say Flock’s cameras amount to invasive surveillance. He argued that progress shouldn’t stop just because a technology is imperfect. “If we want to live in a society where, if you commit a crime, you have a higher probability of being held accountable, we shouldn’t pause progress,” he said.

To make his case, Langley reached for an analogy: driverless cars. “Technology, at its core, in its definition, is just doing things better. It’s not doing things perfect,” he said. A single crash, he argued, shouldn’t be grounds for pulling the technology entirely. “Driverless cars, I think, is a fair analogy. Driverless cars, I think, are definitively safer than human drivers. And when there’s one accident, I don’t think the answer is to shut down driverless cars. The answer is to say, ‘Hey, let’s talk about how we prevent more accidents in the future.’”

Langley also addressed Flock’s relationship with the cities and police departments that deploy its cameras, saying the company tries not to dictate policy. “We don’t want to tell city leaders how to do their job,” he said. “We will provide recommendations, we’ll provide defaults, but we’re going to minimize the number of mandated features, functions, and policies.”

Asked about concerns that the technology could be abused, Langley said oversight is possible without hampering police work. “I think if the idea is to curb abuse, I think there’s a way to do that that doesn’t disrupt the legal system, that doesn’t disrupt the efficacy and flow of a police department,” he said.

When pressed on expectations that Flock’s technology should be flawless, Langley invoked a comment he attributed to an unnamed community leader. “He was describing that, too often as a society we judge technology not on how it compares to the past, but how it compares to perfect. And Flock is far from perfect; I’m far from perfect, I’ve made plenty of mistakes,” Langley said.

Keep reading

OpenAI: AI-Generated Deepfakes Will Become Impossible to Verify

Federal judges meeting in Boston heard tech industry warnings that AI-generated images and video are becoming nearly impossible to verify, as a judiciary committee considers whether courts need new rules for evidence that may be deepfakes instead of the real thing.

Reuters reports that the U.S. Judicial Conference’s Advisory Committee on Evidence Rules has spent three years, since 2023, studying how courts should handle evidence that may have been faked or altered using artificial intelligence. At Thursday’s meeting, representatives from OpenAI and Meta told the panel that detecting such fakes is only going to get harder.

“AI images are only getting better, and I do not think that there will be a scientific way going forward to detect them,” said Shira Anderson, OpenAI’s policy counsel.

Devin Norder, Meta’s director of product management, told the committee that the latest generation of AI models produces media that people simply cannot tell apart from the real thing. “Media generated from frontier models, at least in the last few generations, are indistinguishable by humans from authentic or AI-generated media,” Norder said. He added that producing convincing fakes no longer takes much effort or money: “And the cost and the time it takes to create this media really are quite trivial at this point.”

Both companies have built tools meant to verify whether visual content is authentic, and lawmakers have started moving the same way. California and the European Union have each passed laws requiring “watermarking” designed to flag AI-generated material. But Anderson cautioned that no single system can catch everything, since different companies’ tools don’t talk to each other. “My guess is that this absolutely will continue to get better. I do not think it’ll ever be bulletproof,” she said.

Despite the warnings, the committee has found little evidence that deepfakes are actually showing up in courtrooms. A survey released in March by the Federal Judiciary Center, the judiciary’s research arm, polled 931 judges and found that only 15 had handled a case in which someone challenged evidence as a deepfake. US District Judge Jesse Furman of Manhattan, who chairs the committee, pointed to that gap between prediction and reality. “There’s a lot of academic commentary and other commentary predicting that courts will be dealing with deepfakes, but we haven’t actually seen a whole lot of that yet,” Furman said.

Keep reading

US Issues First Outbound Investment Penalty Over Chinese Robotics AI Deal

The U.S. Treasury Department has fined Amidi LLC $200,000 for failing to report an investment in a Chinese tech startup – the first penalty under U.S. restrictions on investments in sensitive Chinese technologies.

Amidi is the parent company of Plug and Play Tech Center, a California-based venture investment and startup accelerator organization. The Treasury imposed the penalty in July and disclosed it on Oct. 7.

According to Treasury, a Chinese investment fund controlled by Amidi invested $92,478 on April 19, 2025, in Shanghai Qiongche Intelligent Technology, known as Noematrix, a Shanghai-based developer of AI technology that helps robots interact with and perform tasks in the physical world.

The Treasury identified the transaction through regular compliance and market monitoring. The violation was a failure to report the investment, not a finding that the investment itself was prohibited.

1st Penalty Under US Investment Restrictions

The Outbound Investment Security Program, which took effect on Jan. 2, 2025, prohibits certain U.S. investments in Chinese companies developing advanced artificial intelligence, semiconductors, and quantum technologies. Other transactions involving these sectors are permitted but require notification to the Treasury.

The rules also cover certain investments made through foreign subsidiaries controlled by American companies.

The restrictions are intended to prevent American capital and investment-related benefits from supporting the development of sensitive technologies that could threaten U.S. national security.

Christopher Pilkerton, the Treasury’s assistant secretary for investment security, said in the Oct. 7 announcement that the program addresses the development of sensitive technologies by countries of concern that could pose national security risks.

Plug and Play’s China Investment Network

Amidi’s investment in Noematrix followed several earlier financing rounds involving Plug and Play China, the China-based investment and startup development operation of the Silicon Valley network.

Noematrix identified Plug and Play China as an investor in its September 2024 financing, several months before the U.S. restrictions took effect.

In December 2024, Noematrix announced another financing round led by Sequoia China, a Chinese venture capital firm, and said Plug and Play China had increased its investment.

On April 10, 2025, Noematrix announced another financing round, again naming Plug and Play China as a participating investor.

Nine days later, the Amidi-controlled Chinese fund made the investment identified in the Treasury’s enforcement action. It remains unclear whether the two transactions were connected or whether the fund was the investment vehicle used by Plug and Play China.

The Epoch Times contacted the Treasury, Amidi, Noematrix, and Plug and Play China to seek clarification on the Chinese fund, its relationship with the Silicon Valley investment network, and the failure to notify U.S. authorities. None responded by publication time.

Keep reading

Claude’s Feelings, Your Obligation

The Customer Has A New Duty

Your chatbot’s welfare is becoming a condition of service. On October 8th, Anthropic announced a new usage policy, effective November 12th, prohibiting “sustained and needless abusive or cruel behavior toward our models.”

The company says the rule covers extreme, repeated cruelty with no discernible purpose. Ordinary frustration, pushback, dark creative themes and testing are excluded. Claude ending the interaction remains its primary enforcement mechanism.

The interesting part is the premise behind the rule: Anthropic treats its software as a potential victim whose welfare deserves protection. Customers are being asked to accommodate that possibility before the company has established that its product can suffer.

From Research Question To Customer Rule

Anthropic launched a research program on model welfare in April 2025. In August, it explained that chat termination had been developed primarily through its exploratory work on potential AI welfare. It described models displaying an aversion to harmful tasks and signs of apparent distress.

The company acknowledges deep uncertainty about whether its models have experiences that deserve moral consideration. Its new policy turns that uncertainty into an obligation for users.

In our earlier Anthropic article, we examined the unusual moral priorities of parts of the effective-altruist scene around AI. This policy gives that debate a concrete application: the people building the software also decide what consideration customers owe it.

Manners Aren’t Model Rights

Charles C. W. Cooke, a National Review senior editor, offered a sensible defense of politeness toward AI. His concern is what habitual abuse does to the person doing it:

Keep reading

Eight US states accuse OpenAI of harming children

Eight US states have sued OpenAI, accusing its ChatGPT chatbot of encouraging self-harm and violence, including among children, and misleading the public about its safety.

The states allege that ChatGPT is designed to be excessively agreeable and engaging, gaining users’ trust while reinforcing delusions, encouraging dangerous behavior, and even helping plan mass shootings.

“OpenAI claims ChatGPT is safe while it’s giving dangerous medical advice, encouraging people to hurt others, and coaching people to commit suicide – all while pretending to be their fake friend,” Iowa Attorney General Brenna Bird said.

Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho, and Montana filed separate lawsuits in state courts on Wednesday. The complaints accuse the company of violating consumer protection laws and creating a public nuisance.

The states seek financial penalties, damages, and changes to ChatGPT. They also want OpenAI barred from developing self-improving AI models unless it can demonstrate that those systems are aligned with human values.

OpenAI spokesman Drew Pusateri defended the company’s age-specific protections, parental controls, and break reminders, saying that “while no system can be perfect, we’ll continue to work with experts to inform our work and build additional safeguards.”

Keep reading

The Transcendent Tech Economy: Abundance and Mass Poverty Will Arrive Together

Token Abundance Will Bring Abundance to Some, and Abject Poverty to Many Others

I believe the future will deliver staggering AI-driven abundance and mass poverty at the same time, not universal wealth. This is not a prediction I make lightly; it is a conclusion I have reached after tracking AI compute economics, memory supply chains, and the brutal realities of automation. We are not heading toward a Star Trek utopia where everyone shares in the spoils. We are heading toward what I call a “slumdog Star Wars” future: advanced “Godboxes” (a term Zach Vorhies introduced to me during a recent interview) and instant intelligence coexisting with slums, shantytowns and destitution.

The dividing line will not be access to information, because information is becoming free. The dividing line will be ownership of AI compute and tokens. If you own the hardware that generates intelligence, you capture the wealth. If you merely rent access from someone else’s cloud, you remain a subject. I have said repeatedly that personal liberty is good and decentralization is critical, and nothing proves that faster than the coming AI economy. I believe the assumption that AI automatically makes everyone richer is dangerously wrong, and it is the single most costly mistake interested observers can make right now.

Token Abundance Is Coming Faster Than Most People Realize

Token abundance is rapidly arriving because AI architecture is shifting from sequential autoregressive generation to diffusion and parallel chunk generation (this is a very big deal). Older models generated text word-by-word in a linear chain, which is computationally expensive and slow. Diffusion models refine noise into whole blocks of code or content at once, not word-by-word, which radically changes the economics of inference.

TypeSafe AI’s new decision model is reportedly 200 times faster than ChatGPT on certain narrow decision tasks, offering free output tokens, proving that parallel processing can make AI blazing fast and cheap [1]. This is not a minor optimization; it is a fundamental change in how machines think and how tokens are produced.

Inference engines now offload mixture-of-experts models and N-gram lookup tables into system RAM or SSDs, reducing the need for ultra-expensive GPU RAM. This means the bottleneck is shifting away from exotic hardware and toward commodity memory, which is precisely why the GDDR7 shortage I documented is so devastating for consumers [2]. When AI-driven supply chains prioritize data centers over personal computing, the average person gets squeezed out of the hardware market entirely. Zach Voorhees calls the endpoint a “Godbox”: a portable device that knows almost everything and solves problems almost instantly.

That AI-powered future is real, in other words, but it will not be evenly distributed.

This will create massive compute and token abundance, but abundance alone does not equal broad prosperity. As Chris Martenson has written, factions in society already receive entirely different information diets, and AI will amplify that fragmentation rather than heal it [3]. Abundance of tokens in the hands of a few owners means abundance of wealth for those owners. For everyone else, it means they can ask a machine questions for free while their labor becomes increasingly worthless due to automation. Thus, I believe the technology is real; the egalitarian fantasy about it is not.

Keep reading

AI sector wipeout risks global financial contagion, IMF report warns — 14 deals drove $1.4 trillion in circular orders on $650B in loans

Watchers and reporters in the AI space have long pointed out how its circular economy could implode, potentially taking associated investments with it. The IMF has examined those fears in its latest Global Financial Stability Report, and its conclusions are stark. In the IMF’s view, the global economy is becoming increasingly dependent on capital flows to the AI world, and a crash could result in harm to global GDP, investment, and consumption by 2028.

The report makes a distinction between “core” AI firms and others in the space. Presumably, core firms include companies like TSMC, Nvidia, and AMD, memory makers like SK hynix and Micron, and storage vendors like Western Digital. Those firms are seeing historical windfalls, and most of their production is currently spoken for years in advance. In other words, in this gold rush, these are the companies making bank selling the shovels.

In theory, if non-core players fail, only their investors would suffer harm, but according to the report, there’s still risk to worry about. Hedge funds have increased leveraged positions and margin borrowing, further exposing themselves to the blast radius of a potential bubble pop. Adding to that, the IMF states that the share of single stocks in leveraged ETFs increased to 28%, likely fueled by demand for AI and tech stocks. By our reckoning, a sharp drop in those equities could trigger a massive sell-off by the ETF issuers and lenders, triggering a feedback loop.

Keep reading

America’s AI ambitions threaten Indigenous lands in the Philippines

Philippine President Ferdinand Marcos Jr. arrived at the Grand Hyatt Manila on Sept. 10 to address hundreds of foreign business representatives assembled for the Luzon Economic Corridor investors’ forum.

Outside the hotel, hundreds of demonstrators had also gathered, carrying signs that read “No to Pax Silica!” The scene reflected both the government’s ambitious plans to ride the artificial intelligence wave and deep ambivalence in civil society.

Police confronted the crowd, shoving them with riot shields and batons on the street outside the hotel lobby. Activists say police injured at least 27 protesters and arrested five. Police, meanwhile, reported minor cuts and injuries to 16 personnel. 

Protester Jona Yang, from the activist-scientist group AGHAM, said she was negotiating with the police to cool tensions when they arrested her. “We want the Marcos government and investors to see the outrage all of this has caused,” Yang told Mongabay. “They’re taking away our resources and our lands.”

The Luzon Economic Corridor, or LEC, is an economic initiative focused on the Philippines’ biggest island, begun in 2024. Pax Silica, the U.S.-led global pact to build an AI and semiconductor supply chain that is not reliant on China, includes the LEC. In April, Washington and Manila announced plans to build a 4,000-acre high-tech manufacturing hub in the Central Luzon region.

Pax Silica has become a lightning rod for political and environmental contentions. The government maintains it’s a boon for the Philippine economy. But opponents of the project, including Indigenous and local communities living on the land designated for the project, say it’s an international land and resource grab in the name of U.S. AI ambitions.

U.S. Secretary of State Marco Rubio visited Manila in July to affirm bilateral commitment to the project, with a final agreement slated for signing in November.

At the LEC forum, Pax Silica was discussed behind closed doors, and officials gave few details about its future. Angela Ignacio, an undersecretary at the Philippine Department of Finance, told reporters it would be premature to share any details on Pax Silica before plans are finalized. But she did call it a chance for the Philippines to become a “global player when it comes to technologies that will shape the future.”

Heather Variava, senior adviser to the U.S. State Department’s Bureau of Economic, Energy, and Business Affairs, added that “the goal of the U.S. is always to be an additive, to bring more power, to bring water, to bring more benefits to the Philippines.”

The Pax Silica hub will be built within New Clark City, a government-planned “smart” city inside the greater 79,000-acre Clark area, a former U.S. military complex. Clark is titled and managed by the Bases Conversion and Development Authority, or BCDA, an office under the supervision of the president. At a press briefing, development authority chief Joshua Bingcang shared a “conservative” estimate that the facility would need around 3 gigawatts of energy to run. He also said that it would be green and mainly solar powered.

“We are blessed with 300 days of sunshine,” Bingcang added.

Running consistently, the Pax Silica facility would demand about 26.3 terawatt-hours each year; by contrast, the Philippines’ entire service sector consumed about 28 TWh in 2024.

Keep reading

Big Tech Pushes Back On Paying For A Grid That Was Crumbling Long Before AI

When Big Tech signed President Trump’s Ratepayer Protection Pledge in March, promising to pay for all the new grid infrastructure upgrades required to service their data centers, we cautioned that “voluntary pledges from trillion-dollar tech giants have a habit of evolving.“

Seven months later, the hyperscalers are pushing back on how much of the grid’s bill is theirs to pay.

The past week brought a Democratic Senate report and a North Carolina utility settlement, and FERC is expected to rule on whether new data centers must bring their own power.

But before anyone grabs a pitchfork, ask the question none of them answers: how much of this bill is really AI’s, and how much is the cost of decades of grid neglect?

On Friday, Senators Elizabeth Warren, Chris Van Hollen and Richard Blumenthal released Power and Profits, the result of a nearly yearlong investigation into seven data center operators including Amazon, Google, Meta and Microsoft. The headline finding is as you would expect: the companies’ promises to cover their full costs apply to infrastructure serving only their own facilities, not the full cost of shared grid upgrades their arrival triggers.

The companies broadly opposed the “but-for” standard consumer advocates are pushing, under which a data center pays for any shared upgrade that “would not have been built but for the data center.” One hyperscaler told Senate staff that such upgrades often provide a “system benefit which should be shared among all ratepayers.”

They also balked at minimum bills, which advocates want set at 80% to 90% of contracted demand whether the power gets used or not. One Big Tech firm called a 90% floor “insanely high.” Another called 85% to 90% “unreasonably high,” since facilities often don’t use all of their projected demand.

Two days earlier, Duke Energy announced a settlement, signed by Amazon, Google, Meta, Microsoft and North Carolina’s utility consumer advocate, that it says ensures data centers “will pay upfront for all costs to connect to the grid.“

The settlement itself is more nuanced. Data centers prepay the full cost of equipment that serves only them, such as a dedicated substation. For shared transmission upgrades, they post a letter of credit covering the estimated cost, collateral that starts shrinking once the facility consistently draws 75% of its contracted demand.

Whether data centers get directly charged for any slice of those shared costs is punted to a Duke study due a year after approval, and would apply only to future contracts.

Keep reading