 |
|
|
|||||
The sudden wind-down of Anthropic technology within the U.S. government is raising concerns about whether federal officials, without access to Claude, might fall behind in the quest to guard against the threat of AI-generated or AI-assisted nuclear and chemical weapons. Though the rollout has been messyand Claude remains in use in some parts of the governmentthe Trump administrations anti-Anthropic posture could have a chilling effect on collaborations between AI companies and federal agencies, including partnerships focused on critical national security questions related to these kinds of futuristic threats, several sources tell Fast Company. The worry is that severing ties with the company could both limit government researchers understanding of how, in the future, bad actors could use AI to generate new types of nuclear and biological weapons, and hold back scientific progress more broadly. Since at least February 2024, Anthropic has participated in a formal partnership with the National Nuclear Security Administration, the federal agency charged with monitoring the countrys nuclear stockpile. The point of that work, the company has previously said, is to evaluate our AI models for potential nuclear and radiological risks. The concern, here, is that developing nuclear weapons requires specialized knowledge, but that AI, as it continues to advance, could eventually become adept at developing this expertise on its own. Eventually, a large language model might be able to help someone figure out how to design an incredibly dangerous weaponor even come up with a novel one itself. Now, in the wake of President Donald Trumps Truth Social post demanding that federal workers stop using Anthropic tech, its not clear what might happen to Anthropics efforts to guard against these future threats. Some federal agencies appear to still be weighing how to approach the Claude use cases they already have, while others are cutting off access to the tool entirely. As directed by President Trump, the Department of Energy is reviewing all existing contracts and uses of Anthropic technology, a spokesperson for the NNSA tells Fast Company. The Department remains firmly committed to ensuring that the technology we employ serves the public interest, protects Americas energy and national security, and advances our mission. Anthropic declined to comment. Safety concerns at the Energy Department For the past few years, Anthropic has been collaborating with or providing technology to the myriad agencies and national labs that fall under the Department of Energy. For instance, the Lawrence Livermore National Laboratory began using Claude for Enterprise in 2025 and, at the time, made the tool available to about 10,000 scientists. The lab said last year that the technology was supposed to help accelerate its research efforts in the domains of nuclear deterrence, energy security, materials science, and other areas. Anthropic has also worked with the National Nuclear Security Agency on evaluating potential AI-related nuclear safety risks. For example, the agency has provided Anthropic with high-level metrics and guidance that have helped the company analyze the threat of its own technology. Anthropic has also worked with the NNSA on developing technology that can scan and categorize AI chatbot conversations and search for signs that someone might be using an LLM to discuss building a nuclear weapon. An inventory for 2025 for the Department of Energy disclosed that the agency was using Claude at the Pacific Northwest National Laboratory, the Lawrence Livermore National Laboratory, and the Idaho National Laboratory in pilots. Anthropic is also one of several partners in the agencys Genesis mission, which aims to accelerate scientific development by leveraging artificial intelligence. Those collaborations may now be in jeopardy. Claude is everywhere in the Energy Department’s labs, including at the NNSA, according to Ann Dunkin, the departments former chief information officer. If labs, or the NNSA, are working on projects using Anthropic as their AI tool, they are going to have to, at the very least, stop work and start with a new vendor, Durking tells Fast Company. This will cost time and money. More than likely, there will be [new] work as they will have to train a new model. To conduct simulations that involve studying various AI risks, its important to understand how all AI models might behave, she adds. In regard to nuclear weapons, theres worry that AI could be used to gather enough information to build one such weaponor be jailbroken so that it could provide that information, Dunkin says. A former Department of Homeland Security official who focused on AI safety issues echoes those concerns. Anthropic, the person tells Fast Company, was a leader on evaluating how AI models, including its own, might create serious safety risks related to chemical and nuclear weapons. Pressure to remove Anthropic risks wasting peoples time and may not be successful anyway, they say. It also puts federal officials behind on trying to understand the full risks related to artificial intelligence, or to fully benefit from its efficiencies, given that Anthropic is still the leading provider of some AI capabilities. Theres no ban on Claude for the bad guys, the former official adds. Overall, the governments sudden turn against Anthropic risks scaring off other companies that might want to work on serious issues, including those related to nuclear security. Anthropic learned that once youre serving the U.S. government, you might not have the right to say no, at least now without retaliation. Naturally that will deter others from working for the government, especially on sensitive topics, says Steven Adler, an ex-OpenAI employee who focuses on AI safety issues. There’s a bitter irony here: The administration is simultaneously demanding AI companies help with national security and making it harder for responsible actors to do exactly that, Alex Bores, who is running for a House seat in New York on a platform focused on AI regulation, tells Fast Company in a statement. AI companies working with NNSA to evaluate risk isn’t a liabilityit’s a model. Punishing it sends exactly the wrong signal at exactly the wrong time.” An incomplete exit plan Its not immediately clear how federal agencies are supposed to approach Anthropic technology right now. Trump used Truth Social to demand that federal agencies immediately cease all use of Anthropics technology, though such instructions are ordinarily communicated through the federal chief information officer. The Trump administration is reportedly working on an executive order related to Anthropic, while Anthropic has filed a lawsuit challenging its designation as a supply chain risk. The General Services Administration, according to one post, seems to be interpreting the Truth Social post as a national security directive. Te agencys GitHub repository shows that Claude was recently removed for its interagency AI resource, and a person within the agency confirmed that employees could no longer access Claude internally. Still, another person at the agency tells Fast Company that no official instructions about how to actually enforce removing Claude from federal use cases have actually been sent to employees. One major challenge with stripping Anthropics technology from the federal government is that the technology can be delivered in many ways. In Claudes case, this includes products sold by Anthropic directly, but also integrations with popularand controversialgovernment contractors like Palantir and Amazon Web Services. Notably, Claude for Government is still listed as one of the features offered within the Palantir Federal Cloud Service, and several agencies have authorized the use of this package, including the Brookhaven National Lab and the Environmental Management Consolidated Business Center, as well as the State Department and the Treasury. The product describes Claude as purpose built to meet high government security requirements. Palantir also has a long-standing relationship with the NNSA that predates LLMs. The NNSA spokesperson declined to comment on how they were approaching the use of Claude and classified systems. At the time of this writing, Palantir had not respondd to a request for comment. On the military side of government, much ado has been made of the fact that only Caude, and not systems like ChatGPT, has been cleared to operate in classified environments. The Pentagon has since sent a memo to employees that prioritizes removing Claude from any systems that involve nuclear security. Classified environments are also important to civilian agencies. Though Treasury Secretary Scott Bessent has said his agency will be terminating use of Anthropic products and Claude, there was at least some grumbling at a recent meeting focused on AI use within the agency that other AI tools werent similarly available for classified information.
Category:
E-Commerce
For decades, the relationship between a fan and a content franchise was defined by scarcity: You watched a movie or binged a season, then had to wait months or years for the next installment. Fans were spectators, limited to going to the concert, game, or movie, buying merch, and tweeting about their favorite moments. Scarcity trained audiences to wait. Abundance taught them not to. The greatest challenge facing intellectual property (IP) owners today is staying connected and culturally relevant in a world where content is everywhere, all the time, and limitless. While these players historically have been limited by the time and budget required to create premium content, the audiences appetite to relate to their favorite brands has only grown. With the rise of generative AI, media companies can now leverage technology to bridge this gap, transforming how audiences engage with IP and creating opportunities for fan experiences that operate 24/7. Fandom as creation Over the past 10 years, the demographics of fandom have shifted dramatically. According to Ogilvys Fandom Flux study, 86% of Gen Z identify as fans. But for this group of fans, consuming their favorite story doesnt cut it; they want to step inside it, remix it, and co-author the worlds they love. Younger audiences are already making this shift. Gen Alpha is the first generation to experience fandom as an act of creation rather than consumption, according to Ogilvy. As a testament to this trend, 66% of Gen Z and Gen Alpha report spending more time with fan-created content than with official content. Generative AI can solve this discrepancy. By collaborating across industries, IP owners can meet this need, using AI to scale fan engagement in ways that traditional production pipelines cant. Disneys agreement with OpenAI allows fans to generate scenes featuring hundreds of characters within brand-safe parameters. This turns casual Disney fans into active creators. By enabling canon-adjacent storytelling, studios can transform viewers into stakeholders. And when fans can contribute to a world, the franchise grows with its community, and engagement becomes shared rather than one-sided. Gaming already provides a blueprint for whats possible. Fortnite Creative 2.0 has evolved beyond a “battle royale” into a social ecosystem where brands like Star Wars inhabit “living maps, where players can roam, complete challenges, and interact with characters in real time. Now, AI brings that interactive agency to pure storytelling, removing technical barriers so that engaging with a character is as natural as texting a friend. Infinite worlds, infinite engagement As more fans expect this level of immersion with the brands they love, we see a future where AI-powered characters live anywhere, engaging consumers wherever they are. Instead of waiting for the next film or episode, these characters will be available around the clock. Fans can create new moments with their favorite characters, rather than re-watching the same story on repeat. For media companies, AI characters serve as more than just engagement and monetization tools; they are a powerful feedback loop. By partnering with AI platforms, studios can test character dynamics, gauge reactions to new lore, and understand what resonates with their most highly engaged users before greenlighting expensive productions. The business case is undeniable: Platforms that integrate AI-driven personalization are already seeing an average 35% increase in engagement rates, with streaming leaders leveraging these tools as the backbone of their fan retention strategies for younger audiences. Design for participation The future of entertainment is no longer one-way. AI is enabling experiences that are persistent, co-created, and reciprocal. Instead of measuring success by how many people buy a ticket for the latest movie premiere, it will be measured by how deeply audiences engage with the franchise. The next generation of IP will be built by designing immersive experiences, not just content. By embracing partner-driven AI solutions, media leaders can transform viewers into stakeholders. As these fans become true stakeholders, franchises will continue to grow and evolve with their communities, ensuring that today’s favorite characters remain culturally resonant for the next generation of audiences. Karandeep Anand is CEO of Character.AI.
Category:
E-Commerce
The ongoing war in the Middle East continues to embroil new participantsfrom residential properties in Dubai to protestors in Iran getting caught in the crossfire of drones and missiles. And at the same time, global trade is slowing to a crawl, thanks to the effective shutdown of the Hormuz Strait, through which 11% of all global trade passes. Yet another sector finding itself in the firing lineliterallyis data centers. A number of them in the region have been hit by enemy strikes during the two-week war, causing damage and outages. Data centers are an important part of modern economies, enabling the delivery of digital services that keep countries going. Therefore, its little surprise that theyve been targeted by both sides of the war as an attempt to sow chaos and force a capitulation. Data centers are also deeply exposed to wider disruption in the region because they sit at the end of long, fragile supply chains. Many of the chips, memory modules, networking switches, and cooling systems they rely on depend on materials that transit through Middle Eastern choke points or are produced in nearby statesfrom helium and other specialty gases used in semiconductor manufacturing to metals and finished components moving between Asia, Europe, and North America. The nearhalt in shipping through the Strait of Hormuz has pushed up transport costs, squeezed airfreight capacity, and driven insurers to hike warrisk premiums, making it more expensive and slower to move everything from server racks to backup generators and fuel. At the same time, the strait is a critical artery for oil and liquefied natural gas (LNG), so any prolonged disruption feeds directly into higher global energy pricesraising the cost of the vast amounts of electricity and cooling that hyperscale data centers consume, and making new projects harder to finance. Thats less of a problem for the United States, which has its own energy supplies, is the worlds largest LNG exporter, and is insulated from Gulf disruptions by its own abundant domestic production. Data centers already face threats Beyond the immediate impact, theres a corollary risk to the conflict for data centers beyond the Middle East. Abe Silverman, an assistant research scholar at Johns Hopkins Universitys Ralph OConnor Sustainable Energy Institute, says the Middle East conflict isnt primarily a direct supply-chain story for data centers. The biggest threat to data centers isn’t actually oil traffic or disruption to global supply chains, he says. The biggest threat to data centers today is the perception that they are raising costs of electricity for everyday consumers.” Currently, marine traffic through the Strait of Hormuz has practically stopped, including shipments of LNG from the region. If that ongoing disruption continues and pushes up natural gas and electricity prices, consumers may blame data centers for worsening already painful power bills, Silverman believes. While those physical and economic pressures will take months to fully work through supply chains and power markets, the more immediate consequence may be political: As energy prices rise, regulators and communities could increasingly scrutinize whether new data center campuses are worth the extra strain on already expensive electricity bills. We would not anticipate a material shift in companies’ plans and a further expansion in U.S. data centers, but it is a consideration for those focused on Europe and the Middle East, says Julien Dumoulin-Smith, managing director and senior equity analyst at Jefferies, a global investment bank. Theres also the financing of these megaprojects, particularly closer to the center of the conflictand whether its possible for them to be safely insured to be built. Some $2.5 billion of deals to build data centers in the Middle East were brokered last year, according to S&P Global Market Intelligence. If the safety of that infrastructure, and the return on investment, cant be guaranteed as tension in the region continues to ratchet up, it becomes a much harder choice to invest there. That could cause some projects to fall by the waysideor worse, to shift investment in them to states hostile to the West. The impact will be that theyll be rebuilt fairly quickly, and if the Americansand Europeansarent quick off the mark, theyll be rebuilt with Chinese investment, says Lynette Nusbacher, a former Canadian and British army intelligence officer. But beyond that, each new attack sends a message, reckons Nusbacher. Data centers are an important part of the post-petroleum future of the Gulf monarchies, she says. Attacking a data center isnt symbolic, but its a way to show that the U.S. cant guarantee any kind of security for their future.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||