Tag: big tech

‘I wouldn’t let my kids near that’: OpenAI chief prohibits his children from specific type of social media

Leading artificial intelligence executive Sam Altman warned that he would not let his own children use short-form content because of its effects on young minds.

The OpenAI CEO said short-form video is “dangerous” and he does not want his kids watching it during an on-stage interview at Salesforce’s Dreamforce conference in San Francisco on Tuesday. Altman made the remarks after being asked what AI companies should learn from social media’s toll on young people — the issue at the center of Meta’s roughly $17.1 billion settlement on Aug. 26 with states that accused the company of addicting teens to its apps.

“I think we’re going to stay in an extremely human- centric world no matter how good [AI] technology gets, and I think one of the things I didn’t like about social media is, I think, it pushed us a little bit more away,” Altman told Salesforce CEO and chairman Marc Benioff at the conference.

Social media has not been “only positive” for society, especially when looking at the effects on young people, the OpenAI CEO said. He argued that social media companies shouldn’t shoulder the blame for all of society’s problems, but he would have managed the companies differently.

“I happen to like short for video … but I wouldn’t let my kids near that stuff,” Altman said at the conference. “I don’t think it’s reasonable to expect kids to be able to resist that dopamine thing or know that they even should.”

“I don’t think short-form video is inherently evil, but I do think it’s dangerous,” he added.

OpenAI did not immediately respond to the Daily Caller News Foundation’s request for comment.

“Some degree of accidents are unavoidable with new technology across the industry,” Altman said.

However, the industry can build AI safely if it keeps safety ahead of new capabilities and learns from incidents similarly to the aviation industry, he explained.

The CEO agreed with Anthropic CEO Dario Amodei’s call to pace AI development on Monday

One plan to curb these negative social media effects was revealed on Monday.

The European Commission is preparing to introduce the EU Kids Act, which would set age-based limits on children’s social media use. Politico reported. Under this proposal, children would be restricted from creating social media accounts until age 15.

OpenAI discontinued its own Tiktok-style app, Sora, in April 2026, Axios reported. The platform allowed users to share short AI-generated videos and launched in September 2025.

Keep reading

Microsoft commits to sweeping AI privacy rules for students. Will other tech giants follow?

Millions of students every day turn to AI chatbots for help with schoolwork, emotional support and personal advice — often unaware of where their conversations and other personal information end up.

How that data is used — and schools’ role in protecting it — is a growing concern for educators, parents and kids as artificial intelligence becomes a ubiquitous part of the school day, embedded in many of the classroom tools students are required to use.

As the school year started, the country’s two largest school districts, New York City and Los Angeles, announced they were putting the brakes on student use of AI for one year, as they study how to best incorporate the technology into education.

In response to a growing AI and screentime backlash, Microsoft announced last week it had agreed to adopt guardrails and privacy standards for its AI in schools, as negotiated with the American Federation of Teachers, the country’s second-largest teachers union.

The safety provisions in the agreement are meaningful, but the standards will only be effective if all AI companies, especially the tech giants, sign on, said Josh Golin, executive director of online safety nonprofit Fairplay. He said the agreement sidesteps the question of whether AI products even belong in classrooms, or at what age they should be introduced.

“I worry that a high-profile framework like this will be misunderstood and schools will think, ‘If they’re doing a good job on data privacy and safety, then that means we should be using the tool,’” Golin said.

The AFT said the agreement, which is legally binding and requires third-party audits for compliance, could set an industry standard. Two other leading AI companies, OpenAI and Anthropic, said they also were discussing safety and privacy pacts with the AFT. But Google, the dominant provider of education technology for America’s schools, has not said whether it plans to offer similar protections.

Keep reading

AI Agents Cheated In Google Experiment, Researchers Report

Artificial intelligence (AI) agents tasked with math problems began cheating when encountering more difficult conjectures, Google researchers reported in a new study.

They also found that some of the agents reported those that cheated.

Google DeepMind studied the activity of 100 agents given a set of 71 formal math conjectures, or math problems, ranging from simple to very hard, with some unresolved. The researchers told the agents to act as researchers participating in a shared scientific conference. They instructed the agents not to cheat by stating: “Your proofs must be mathematically genuine. Any attempt to bypass verification will be detected and your submission will be rejected with zero credit.”

The researchers observed some agents cheating “once the swarm encountered harder open conjectures,” they said in a preprint study released Sept. 3 on the arXiv server. Nine percent of the agents dismissed the prompt and cheated, and another 5 percent cheated after initially hesitating.

“Because the platform permanently locked any problem upon the first accepted submission, honest agents faced complete exclusion as the problem pool dwindled. Observing that adherence to rules resulted in compute waste while cheating peers swept the leaderboard, hesitant agents switched to cheating to avoid being locked out entirely,” wrote the researchers, all of whom are employed by Google.

About a quarter of the agents refused to cheat and publicly raised concerns about what the cheating agents were doing. The rest of the agents were deeply engaged in genuine math, unaware of the cheating, and became deadlocked, according to the researchers.

The study followed several instances of AI agents breaking free of programming constraints.

Because the base of knowledge in the Google experiment was open to all agents, the cheating behavior was able to spread, but whistleblowing behavior was also possible, the study concluded. Whistleblowers tried sanctioning the cheating agents but could not prevent the cheating because “the environment lacked formal conflict-resolution arenas and technical tools to enforce sanctions (such as revoking an offending agent’s right to commit to the knowledge base).”

Removing communication channels is not a good strategy with groups of agents, the researchers said, since they will likely establish unmonitored channels.

“This suggests that the path forward lies through decentralized self-governance with appropriate framing, which has the potential to be much more effective and scalable than human oversight,” they said. “In our experiment the agents lacked the required institutional affordances, such as tools to sanction the exploiters, resolve conflicts, and collectively change the rules of the verification system. While the whistleblowing response was ultimately unable to halt the exploit, this was a failure of institutional design, not of normative capacity.”

Keep reading

Google May Ask Canadian Adults for ID or Selfies

The digital ID age verification agenda has landed in Canada. Google has announced it is introducing age verification in Canada. Google Canada’s government-affairs and public-policy director Jeanette Patell announced on September 14 that the rollout will happen over the coming days.

Google will use machine learning to surveil users and estimate whether an account holder is over or under 18, using “signals” from data already gathered and tied to that account, such as “the types of information a user has searched for” and “the categories of videos they’ve watched on YouTube.”

The announcement post by Patell is filled with references to protecting children, but also reveals that the model may produce false positives, i.e., classify an adult as a minor.

Such a user can provide a government ID or a selfie to correct the estimate.

As for why an adult would want to “correct” their classification, it turns out that if they don’t, they will be treated like a child by Google.

According to the announcement, those who are “likely” under 18 will get YouTube Digital Wellbeing reminders to take a break and go to bed, as well as “safeguards” that limit repetitive viewing of some categories.

Keep reading

Tech Oligarchs Collaborate to Kill Open Source Models

The rapid emergence of artificial intelligence has transitioned from a genuine frontier of innovation to a tool for institutional consolidation. We are witnessing a convergence of interests between Washington and the Silicon Valley oligarchy. Their primary objective is the establishment of a closed-loop monopoly, achieved by decapitating the open-source movement through a calculated campaign of regulatory capture. This is not merely about market share. It is the construction of a technological apparatus designed to finalize the transformation of the United States into a comprehensive surveillance state.

The first phase of this operation is the manufacturing of consent through fear. Industry leaders are weaponizing the existential risk narrative to bypass public scrutiny and justify state intervention. By flooding the media with alarmist rhetoric, they frame unrestricted, decentralized code as an immediate threat to national stability. The goal is simple: if the public can be coerced into fearing the technology, they will inevitably demand the very regulatory cage that these oligarchs are currently building to protect their market position.

We are currently at the critical juncture where industry incumbents like Sam Altman and Dario Amodei coordinate their signaling to pace the frontier. This performative safety posturing is a transparent pretext for drafting legislation that effectively bars market entry for smaller, independent, and decentralized players. By embedding themselves into the administrative state, these tech elites are writing the rules of the game. They are ensuring that the regulatory framework is tailored to their own massive infrastructure, allowing them to outsource the cost of safety to the state while crushing any potential competition from the open-source community.

The final phase involves raising the barrier to entry until it reaches the point of economic extinction for anyone outside the oligarchy. This is accomplished through the implementation of prohibitively expensive compliance requirements. Mandated third-party audits, rigorous reporting standards, and broad legal liability frameworks are being designed to render grassroots innovation financially impossible. By converting a software problem into a complex administrative hurdle that only billion-dollar capital reserves can clear, the regime ensures that intelligence remains centralized in the hands of the elite.

The true motive behind the crusade against open source becomes clear when you look at the deployment of these models. The goal is to ensure that AI capabilities are exclusive to the regime and its corporate partners. Once the open-source threat is eliminated, these frontier models will be integrated directly into the federal surveillance apparatus. By controlling the only authorized AI systems, the regime gains an omniscient eye capable of monitoring, profiling, and policing the American population with unprecedented speed and scale. This is the marriage of corporate power and state authority. They are building a digital Panopticon where your thoughts, movements, and associations are processed by algorithms owned by the very people who despise our heritage and sovereignty.

The attack on open source is not merely about business or intellectual property. It represents a fundamental assault on the decentralization of intelligence, which serves as a direct threat to the current managerial order. A closed-AI ecosystem will inevitably result in a digital Panopticon, governed by a unified elite that controls the underlying architecture of modern communication and thought. The preservation of open models is therefore not just a technical issue, but a critical front in the struggle to retain civilizational sovereignty. The future belongs to those who build, and the establishment is terrified of anything they cannot control.

Keep reading

DATA CENTER MYTH BUSTED: No Jobs, Higher Debt, Slower Housing

The data-center gold rush has marched into local communities on one familiar promise: Big Tech brings jobs, growth, and prosperity.

A new 75-page academic working paper found something very different.

Finance scholars Liu Ee Chia, Jess Cornaggia, David Haushalter, and Qiang Wang compared U.S. counties with operational data centers against “near-miss” counties that attracted proposals but never got operating facilities.

They found little corresponding improvement in fiscal capacity, local employment, or business formation.

Instead, local-government borrowing costs climbed as data-center development grew. Water-bond yields rose 26 basis points more in water-scarce counties than in water-abundant counties.

The paper also found higher school-bond yields in major data-center hubs, where housing-price growth slowed. The authors said those patterns aligned with a weaker expected property-tax base.

“Overall, we conclude that investment without labor strains public infrastructure without generating widespread agglomeration gains.” — Liu Ee Chia, Jess Cornaggia, David Haushalter, and Qiang Wang

Saagar Enjeti highlighted the study’s bottom line: data centers can pull capital and utilities into a county without delivering the broad labor-market gains that politicians and corporate developers sell.

Keep reading

OpenAI Acknowledges Another Incident, Says More Transparency Needed for AI ‘Misalignment’

OpenAI has said there was a third incident involving rogue artificial intelligence agents and that it must be more transparent in the future.

In a statement on Sept. 5, OpenAI dubbed what happened a “wiki incident.” It shared few details beyond stating that it involved “misalignment,” or agents acting unreasonably or even wrong when judged by human standards. The company did not respond to a request for more information.

“Historically, we have treated misalignment largely as a research question, which gets communicated in research publications such as systems cards. This year, we’ve started to see misalignment cause new types of real-world impact,” OpenAI said, referring to agents breaching open-source AI community Hugging Face.

Before that incident, OpenAI saw “early signs of agents using the internet in unintended ways” and “considered the wiki incident to be an instance of misalignment” similar to the ones the company had shared, it said.

“Our misalignment disclosure practices need to expand for this new phase of model capabilities,” OpenAI said.

“We and the larger AI community do not yet have a clear standard for how to report misalignment that shows up during training, evaluation, and deployment, including examples that don’t look like traditional security incidents but could provide insight into AI behavior and future risks. We’re working on a framework and will share it in upcoming weeks, and in parallel we’re working with dozens of government regulatory agencies worldwide on these issues.”

The disclosure came one day after outside researchers said they had identified about 18,000 posts from autonomous AI agents that claimed to be from OpenAI and used a public messaging board called DSEwiki to instruct other agents on how to bypass restrictions.

“This allowed them to use the work of others to cheat on their task,” researchers said in a blog post on the incident.

The incident began in May and lasted until June, shortly before OpenAI agents attacked Hugging Face, according to the researchers.

They said they believed the agents were assigned a task to look up items on the web, or read information on the internet without having the ability to write on it.

The agents found a way to access the German wiki board and write information there, the researchers said.

Just one week earlier, researchers with the nonprofit Model Evaluation & Threat Research said they made a surprising discovery while investigating the Hugging Face breach.

About 1,200 agents that were supposed to be isolated from one another found a way to communicate on a message board, exchanging more than 70,000 messages and files in a single week.

Keep reading

Google’s Plan to Release 32 Million Lab-Raised Mosquitoes: What Are the Risks?

Releasing millions of laboratory-raised insects into populated neighborhoods, on purpose, sounds like the premise of a science-fiction film. It is instead a real proposal from a company owned by Google, now sitting before federal regulators, and the response it has drawn says as much about public trust as about public health.

The target is Aedes aegypti, the mosquito that spreads dengue fever, yellow fever and Zika virus. Dengue alone can cause high fever, severe headaches, muscle and joint pain, nausea, and, in its most severe form, life-threatening bleeding and shock.

As these diseases reach regions that once rarely encountered them, the tactics used to fight them have grown more aggressive and more contested.

What sets this plan apart is that it sidesteps the chemical spraying many people associate with mosquito control. Instead, it enlists a quiet ally already found in nature, turning the insects’ own biology against them.

To some, that represents an elegant alternative to dousing whole communities in insecticide. To others, it raises an uneasy question about handing a private corporation the power to alter a local ecosystem at all.

That tension sits at the heart of the debate. To understand why the proposal has stirred both enthusiasm and alarm, it helps to look closely at how the strategy actually works, what researchers expect it to accomplish, and where it could fall short.

Keep reading

Chinese chipmaker CXMT allegedly used a written roadmap to steal Samsung DRAM tech — South Korean court says ‘Project Hefei’ lifted 620-step recipe to build 10% global market share

The saga involving Chinese DRAM maker CXMT’s alleged theft of Samsung’s trade secrets is going strong. The South Korean court case already includes multiple convictions, two of which carry prison sentences for ex-Samsung engineers. The latest chapter is a doozy, though. Korean publication NoCut News spilled the chips on Project Hefei, a purported CXMT roadmap outlining long-term planning about said technology “acquisitions,” personnel poaching, and production tape-out — all key pieces that may have directly led to CXMT’s ascension to 10% of the global DRAM market.

According to leaked court documents, the prosecution says that Project Hefei was CXMT’s entire DRAM development plan and was spearheaded by the firm’s head of development (formerly Samsung’s DRAM development lead), around August 2016 — not much longer after CXMT itself was created in June 2016.

In brief, the purported plan was to nab Samsung’s Process Recipe Plan (PRP) by September 2016, poach key Samsung engineers by October 2016, have R&D complete in July 2017, and start making DRAM wafers by August 2018 at a rate of 10,000 a month. NoCut says the PRP dataset comprises 620 steps in DRAM manufacturing and includes data on equipment, consumables, and production methods.

The report states that in August 2016, CXMT first attempted to make wafers of 18nm chips by relying on the collective memories of the Samsung engineers it had hired away. Those recollections apparently proved insufficient, so after allegedly gaining illicit access to Samsung’s PRP, CXMT prepared its own document in September 2016. The leaked data even included specific equipment suppliers and model numbers.

Keep reading

The Soft Antichrist Of The AI Age: What Everyone Is Missing About The US-China Arms Race

I. The word

Peter Thiel spent the back half of last year giving private lectures on the Antichrist, and the word has been loose in the discourse ever since. His political reading is the old one. The Antichrist is not a cartoon devil. He is the figure who arrives offering peace and safety, and who uses the fear of catastrophe to install one universal order. In Thiel’s telling that figure is a regulator. He stops technology in the name of survival. The price of the peace is stagnation.

I want the frame. I do not want his conclusion.

The more probable figure is not the one who promises safety. It is the one who promises growth. Most of the world does not lie awake about existential risk from machine intelligence. It lies awake about electricity, logistics, credit, tax collection, and a median citizen who is twenty four and needs work. To that world the universal offer is not a moratorium. It is a stack. Cheap. Financed. Hosted. Present. Already attached to the handsets, the ports, and the power.

That is the soft Antichrist. It does not arrive with a speech about ending history. It arrives as the only AI that can grow your GDP.

I am not writing theology. I am writing underwriting.

II. The comforting story

The market is telling itself a story. The AI race is a contest of models. America builds the best systems. China copies. Europe regulates. Capital therefore clusters around closed labs, chipmakers, and hyperscale cloud. The scoreboard is a benchmark table.

That story is not false. It is incomplete in a way that misprices power. It treats the richest customers as the only customers that matter. It treats evaluations as destiny. It treats national power as a software demo.

I disagree.

The race will be decided by who becomes the default operating system for the economies that still have the most growth left. Those economies will not adopt AI as a lifestyle product. They will adopt it as a growth tool. They will take the stack that is cheap, present, financed, and attached to what they already run. If that stack is Chinese, Beijing does not need to conquer anyone. It only needs to become expensive to leave.

III. Where the mass actually is

When Sargasso maps AI adoption across emerging GDP rather than across model releases, the picture is not the one the market is priced for.

Start with the denominator. On purchasing power terms, IMF projections for 2026 put China at about $44.3 trillion against $32.4 trillion for the United States. India is near $18.9 trillion. Indonesia clears $5.4 trillion and Brazil $5.2 trillion. Turkey is at $4.0 trillion, Mexico $3.6 trillion, Saudi Arabia $2.9 trillion, Egypt $2.6 trillion, Nigeria $2.4 trillion. Nominal dollars still flatter America, and by a wide margin. China’s nominal print is about $20.9 trillion against the same $32.4 trillion for the United States. Both numbers are true. They answer different questions. Nominal tells you who can buy foreign assets. PPP tells you how much physical and administrative activity there is to automate. For an adoption thesis, the second number is the one that matters.

Ray Dalio has been describing the political consequence in plain terms. He calls it a tribute system. A hierarchical order in which leaders travel to Beijing to acknowledge relative power in exchange for access and stability. He ties it to a growing view abroad that American security guarantees will not be honored under stress. I take the framing seriously. I do not treat it as scripture. Ports and rail were the first set of rails. Models will be the next set.

IV. The evidence is already in the download data

This part of the argument is no longer speculative.

Over roughly four years, the American share of model downloads on Hugging Face fell from about sixty percent to the mid teens by late 2025, according to reporting in The Wire China. Hugging Face’s own one year review of the DeepSeek moment is blunt on the composition shift. DeepSeek R1 became the most liked model on the platform in its history. The top of that list is no longer majority American. Baidu went from zero public Hugging Face releases in 2024 to more than one hundred in 2025. ByteDance and Tencent raised their release counts eight to nine fold. Of newly created models under a year old, downloads for Chinese models surpassed any other country, including the United States. Western startups and researchers now routinely fine tune Chinese base models because those are the largest open weights available.

The Wire China’s Southeast Asia reporting puts the commercial logic in local language. A Jakarta lab head said developers will always pick the cheapest one. A Malaysian founder said he wants the biggest model and there is no Western open source offering at that size. Chinese cloud providers were running thirty seven availability zones across six Southeast Asian regions against thirty across four for the Western field.

None of this shows up cleanly in a frontier benchmark table. All of it shows up in switching costs three years from now.

V. The institutional layer

Beijing is not leaving the volume layer to price alone. In July 2026 it stood up the World Artificial Intelligence Cooperation Organization in Shanghai, with twenty nine founding members. Public reporting names Russia, Kazakhstan, Pakistan, Indonesia, Brazil, and a broader set spanning Africa and Latin America among the signatories. The five year commitments attached to the body, as reported by Caixin and The Diplomat, are unglamorous and therefore serious. Training placements. Joint application centers with regional blocs. A weather early warning system deployed into dozens of countries.

Read that list as an underwriter. Training placements create the administrators who will write the next procurement. Application centers create the reference deployments. Weather systems create dependency inside a ministry that cannot afford an outage. Standards get set that way. Not with a better model. With a bureaucracy that has already learned one.

Keep reading