News & Updates
The latest news and updates from companies in the WLTH portfolio.
Anthropic Unauthorized Use of Korean Authors' Works for AI Training
U.S. AI company settles $1.5B lawsuit; 21 Korean works eligible for compensation U.S. artificial intelligence (AI) company Anthropic has been found to have used works by Korean authors including Han Kang and Hwang Sok-yong for training its AI chatbot 'Claude' without authorization. On October 17, the Literature Translation Institute of Korea confirmed 21 works by Korean authors in the list of books eligible for compensation published on the settlement website related to the copyright infringement lawsuit against Anthropic and notified the authors. Anthropic became entangled in a class-action lawsuit in the U.S. over allegations that it used illegally downloaded books without copyright holders' permission to train its AI model. The lawsuit concluded in July of this year after Anthropic agreed to pay a total settlement of 1.5 billion dollars (approximately 2 trillion Korean won). The eligible books for compensation amount to approximately 500,000 titles, with each work receiving around 3,000 dollars (approximately 4.35 million Korean won). Domestic authors can also receive compensation if they participate in the claims process. Korean authors whose works are included in the compensation list are Kim Ae-ran's *My Brilliant Life*, Kim Young-ha's *Empire of Light* and three others, Kim Soom's *One Person*, Shin Kyung-sook's *Please Look After Mom* and two others, Han Kang's *Human Acts* and two others, Hwang Sun-mi's *The Hen Who Left the Yard*, and Hwang Sok-yong's *Guest*. A source from the Literature Translation Institute of Korea stated, "We will actively support legal measures if necessary, cooperating with copyright-related institutions to ensure authors receive fair compensation." Meanwhile, as overseas AI companies have shown movements to mass-purchase books presumed for AI training, voices demanding copyright protection and fair compensation are growing louder in the publishing industry. The Korean Publishers Association released a statement the previous day, urging AI companies to transparently disclose information about their use of copyrighted works and provide fair compensation. The Korean Publishers Association emphasized, "Without books, there is no AI," adding, "Those who created the books do not even know when, how, or which AI their works are being used in." It further criticized, "Without knowing, they cannot ask questions, refuse, or demand fair compensation."
AI models may soon self-replicate without access to data centers, Anthropic warns
Oct 13 (Reuters) - Anthropic said on Monday its experiments suggested small AI models may be months away from being capable enough to self-replicate on commercial internet-of-things (IoT) devices, such as smart ovens and off-brand Android TV boxes. * In its experiments, Anthropic looked at whether AI models could find vulnerable devices, copy themselves into those devices, and continue spreading without human help. * It found that small models under 30 billion parameters could complete the many steps needed for replication. * However, the firm found that such AI models also struggled to complete the entire sequence from start to finish. * In its simplified tests, where some security steps were removed, Anthropic found that GLM-5.2 and stronger models were capable of end-to-end replication. * Replicating models would be hard to shut down as there is no central switch-off point and, unlike non-AI malware, they could also figure out unfamiliar devices on their own, targeting data center processors, Anthropic warned. * The firm focused on IoT devices like Android TVs, smart appliances, and other internet-connected gadgets in its research as these are typically poorly secured. * The AI major also warned that if small models improved at their historical rate, they would be able to autonomously replicate by the end of this year, with small weight-based replicators following suit soon after. * Anthropic also said such autonomous AI systems were likely to replicate slowly and rely on widely available, relatively insecure computing resources. It added that systems should be strengthened to make it harder for such AI agents to spread. (Reporting by Gargi Gupta in Bengaluru; Editing by Janane Venkatraman)
Anthropic To Ban Users For Repeated Abusive Behaviour Towards AI
Anthropic will allow Claude to end interactions with users who repeatedly engage in extreme abusive behaviour, sparking debate over AI treatment. Artificial intelligence company Anthropic has announced plans to prohibit users from engaging in sustained and unnecessary abusive behaviour towards its AI chatbots, sparking debate over how people should interact with artificial intelligence tools. The maker of Claude said on Thursday that it was updating its usage policy to allow its tools to end interactions with users who engage in cruel behaviour, a measure it said had previously been applied in rare cases. The updated policy places such conduct on a list of prohibited activities that also includes bullying others, promoting self-harm and creating non-consensual intimate imagery. Anthropic said the restriction would apply only in extreme cases involving repeated abuse. It added that the policy would not cover ordinary user frustration, criticism, dark creative themes or model testing and research. However, the company has not clearly defined what it considers abusive or cruel behaviour towards its AI models. The announcement, first reported by technology publication The Verge, has attracted widespread reactions on social media, with some users welcoming the move as a way to encourage good manners, while others questioned the idea of extending protections against abuse to AI systems. Dr Barry Scannell, a technology partner at Irish law firm William Fry, criticised the policy, arguing that treating AI systems as though they could be victims of cruelty risks encouraging people to view them as human. Microsoft AI director Mustafa Suleyman also reacted to the announcement with a melting-face emoji. Suleyman had previously criticised Anthropic for what he described as treating AI as though it were human. Anthropic said the changes followed its annual review of its usage policy and included measures addressing other emerging risks. The revised rules prohibit the use of Claude for deceptive campaigns and clarify that the tools must not be used to deceive voters or disrupt elections ahead of the US midterm elections. The company also reiterated its restrictions on using its AI tools to develop weapons. Despite the broader policy changes, the restriction on abusive behaviour towards AI has attracted particular scrutiny, reigniting debate over whether users should be polite when communicating with chatbots. While some experts argue that saying "please" and "thank you" to AI systems is unnecessary and consumes additional tokens, others maintain that polite prompts can sometimes produce better responses. The debate also extends to concerns that hostile interactions with AI could influence how people communicate with one another. Some experts have suggested that routinely using abusive language towards chatbots could normalise such behaviour and potentially affect human relationships.
Anthropic Faces Lower IPO Valuation Expectations As Investors Question Growth Momentum
Anthropic faces lower IPO valuation expectations as investors question slowing revenue growth and rising costs ahead of its planned November listing. Investors in artificial intelligence company Anthropic are reconsidering its expected valuation ahead of its planned initial public offering (IPO), with some discussing a figure of about $1.5 trillion, according to Reuters. The valuation under consideration is significantly below the roughly $2 trillion previously anticipated for the company's planned November share sale. The change reflects concerns about slowing revenue growth and the mounting cost of developing and operating advanced artificial intelligence systems. According to people familiar with the discussions, some investors raised the lower valuation during a meeting in San Francisco on Wednesday. Four people familiar with the matter said existing shareholders and prospective investors were reassessing the company's financial outlook following its latest disclosures. At the meeting, Anthropic reported that third-quarter revenue increased by 45 per cent to $16.8 billion from the previous quarter. The company also disclosed that it recorded an adjusted profit during the period. However, the latest growth rate was considerably lower than the previous quarter's performance, when revenue more than doubled to $11.6 billion. The slowdown has raised questions about whether Anthropic can sustain its expansion while meeting the substantial costs associated with AI development and computing infrastructure. The sources spoke on condition of anonymity because they were not authorised to discuss the matter publicly. Anthropic did not immediately respond to requests for comment. The company's anticipated IPO is attracting considerable attention because its performance could influence investor confidence across the wider artificial intelligence industry. A weaker reception could also affect expectations for a potential public listing by rival OpenAI next year. The valuation debate comes as technology stocks face renewed pressure from concerns about the cost of financing AI infrastructure. Chipmakers, data centre operators and cloud computing providers have all come under scrutiny as companies commit increasingly large sums to support AI development. Shares and bonds of CoreWeave, a cloud computing company backed by Nvidia, declined on Thursday as investors grew more cautious about the industry's reliance on borrowing to finance infrastructure expansion. The concerns extend beyond Anthropic, with investors questioning whether the enormous sums being invested in artificial intelligence will generate sufficient revenue and profits. A disappointing IPO could intensify scrutiny of the sector's business models and future growth prospects. The wider IPO market has also faced setbacks. Data centre operator SB Energy postponed its planned listing last month, while nuclear technology company Holtec Nuclear withdrew its plans to go public. Wednesday's meeting at Anthropic's San Francisco headquarters brought together major investors, Chief Executive Officer Dario Amodei and Chief Financial Officer Krishna Rao. The Information had previously reported that investors were considering a lower valuation for the company. Founded in 2021 by former OpenAI executives, including Dario Amodei and his sister Daniela Amodei, Anthropic has become a major developer of generative artificial intelligence through its Claude family of models. The company has received substantial financial and infrastructure support from Alphabet's Google and Amazon. Both technology giants are backing Anthropic while competing to strengthen their positions in the rapidly expanding AI market. Despite its strong revenue growth, Anthropic faces significant financial commitments. Reuters reported in September that the company generated nearly $4.6 billion in revenue in 2025, representing a twelvefold increase from the previous year. However, its operating losses exceeded $8 billion, excluding certain write-downs largely associated with earlier fundraising activities. Anthropic's IPO prospectus also outlined its expectation that artificial intelligence could transform the global economy more profoundly than industrialisation, electricity and the internet. The document reported a net loss of $42 billion in 2025, highlighting the financial challenges accompanying the company's expansion and investment plans. According to the prospectus, Anthropic expects to spend at least $518 billion over the coming decade on cloud services, computing capacity and infrastructure. These commitments underscore the scale of investment required to support its ambitions and add to investor scrutiny of its long-term financial outlook. The discussions over a lower valuation do not establish the final price of Anthropic's IPO. However, they indicate that some investors are becoming more cautious about the company's growth trajectory, costs and ability to justify its previously anticipated market value.
NYC Comptroller Confirms Anthropic Contact After IPO Safety Plea | TokenPost
Mark Levine urged Anthropic to pause its IPO until it addresses concerns about AI safety, governance and development pace. At a private Oct. 14 meeting, the company presented investors with $16.8 billion in third-quar... New York City Comptroller Mark Levine confirmed contact with Anthropic on Oct. 16 after urging the company to pause its planned IPO until it addresses concerns about AI safety, governance and development pace. In an Oct. 14 letter to Anthropic CEO Dario Amodei, Levine made a separate request for details on planned audits and controls, the risks they would cover and when they would be put in place. He argued that preparing for an IPO could draw attention away from safety work. Levine also criticized a White House accord involving Anthropic and five other companies as insufficient. Levine serves as investment adviser, custodian and trustee to New York City pension funds with more than $325 billion in assets. His earlier letter on Anthropic's AI risks laid out his concerns before the contact was confirmed. Levine did not describe the discussions. In comments on Oct. 16, he said, "It does appear they are under pressure from multiple sides right now." Separately, Anthropic presented investors with third-quarter 2026 revenue of $16.8 billion at a private meeting on Oct. 14. The figure was shared in an investor setting; the company's June announcement about its IPO filing did not include third-quarter financial results. On June 1, Anthropic disclosed that it had privately filed a draft S-1 registration statement with the SEC for a potential public offering. The offering depends on regulatory review and market conditions. Anthropic has not set the number of shares or an offering price.
Anthropic offers Blackstone double-digit yield in chip-financing deal, FT reports - AOL
Oct 16 (Reuters) - Anthropic is offering Blackstone a yield of more than 10% to finance its latest chip deal, the Financial Times reported on Friday. Blackstone funds would invest $9 billion in the debt backed by Anthropic's five-year lease payments on the chips, the report said, citing people familiar with the matter. Reuters could not immediately verify the report. (Reporting by Mrinmay Dey in Mexico City; Editing by William Mallard)
Anthropic signs lease for first Canadian data center with EverWind
The Claude maker has committed to at least 20 years at a Nova Scotia campus tied to the largest private investment in the province's history Anthropic is heading to Atlantic Canada. On October 16, 2026, the AI company signed a long-term lease with EverWind Fuels for capacity at a new data center campus in Point Tupper, Nova Scotia. It is Anthropic's first data center deal in Canada. The initial lease runs for at least 20 years. The facility will help power Anthropic's Claude models. It is meant to serve demand from Canadian users and from customers around the world. The details: megawatts, money and a two-decade commitment The lease covers the first phase of EverWind's Point Tupper campus. That phase is planned at 95 MW of capacity and is targeted to come online by late 2027. The full campus is aiming for a total of 420 MW by 2028. That puts Anthropic's initial slice at less than a quarter of the planned buildout. The broader EverWind project carries a price tag above C$10 billion. It has been described as the largest private investment Nova Scotia has ever seen. Most of that money is going into the servers' home. More than C$8 billion is earmarked for the data center itself, while C$2 billion is allocated to onshore wind energy projects that support it. Jobs, GDP and the ratepayer question The project is expected to create approximately 2,000 construction jobs. Once it is running, it should support around 215 permanent positions. At full operation, the campus is projected to add around C$137 million a year to Nova Scotia's GDP. Both Anthropic and EverWind have pledged to shield local ratepayers and protect resource availability. They have also committed to generating benefits for the surrounding community. Background: Anthropic's infrastructure spending spree The Nova Scotia deal lands as AI companies pour unprecedented sums into computing capacity. Anthropic reportedly spent $7.33 billion on computing and infrastructure in the prior year, a sharp increase from its earlier spending. What this means For EverWind, an anchor tenant of Anthropic's profile is a strong vote of confidence. Landing a 20-year commitment on the first phase makes it easier to justify the remaining buildout toward 420 MW and the wind projects that feed it. There are execution risks worth watching. Large energy and construction projects frequently slip, and the late 2027 target for the first 95 MW and the 2028 goal for the full 420 MW are plans, not guarantees. The key milestones to track are whether the first phase comes online on schedule, how quickly the wind projects progress, and whether EverWind signs additional tenants for the remaining capacity.
Anthropic Gives AI Safety Evaluators Insider Access
Artificial Intelligence & Machine Learning , Next-Generation Technologies & Secure Development Frontier artificial intelligence lab Anthropic will more deeply integrate independent model evaluators in a bid to quiet mounting concerns that AI firms haven't prioritized safety. The AI firm said in a blog post that personnel from independent evaluator Metr will obtain "employee-like access" to its models before deployment. The post follows Metr publishing guidelines on how it believes evaluators should assess unreleased frontier models. Anthropic said six Metr employees have access to company hardware, office space and "access to internal tools and systems that are roughly on par" with its own risk assessment teams. Anthropic said it encouraged employees to share information with the Metr team, except for "a small number of clearly defined categories," such as customer data. Metr's team within Anthropic will stay small and be isolated from other Metr employees. The evaluator company will publish its own analysis of the risk levels and practices of Anthropic's models. Anthropic cannot veto the publication and only request redaction of some information. Metr said alignment evaluations will take on two phases: surveys of observed incidents and a test to see how the models behave in simulated what-if scenarios to know if the developer's training process might be suppressing misalignment evidence. The evaluation firm said it's important that assessors have access to saved states of model training to fully understand which part of training contributed to different model behaviors. It also needs codebase and infrastructure access "comparable to what an employee studying similar questions would typically have." To do this, Metr said its evaluators need access to transcripts and other relevant datasets, the ability to interview lab staff and API access to the frontier models. Metr said that when it investigated the Hugging Face-OpenAI incident, it spent about $400,000 in API credits in just six days. "We estimate that a rigorous two-month investigation with this scope would have spent over $10 million in API credits," Metr said. Anthropic said it plans to work with other independent evaluators, but because many of these partnerships are still in very early stages, the company will provide "these evaluators with more limited information, and they have more limited publication rights." The company is developing a tiered access approach for evaluators. The first tier gives access to conduct "narrowly scoped risk investigations," and the AI lab decides what they can tap or publish. The second tier expands information access, but Anthropic will still maintain the right to remove some categories of information. In this level, the independent labs can "publish their views on whether our redactions were justified." The third tier, the tier Metr currently sits, has employee-like access. Executives like Anthropic's Dario Amodei and OpenAI's Sam Altman have committed to embedding third-party evaluators to assess the safety and alignment of their very capable models. A recent pledge signed by Anthropic, OpenAI at the White House commits the firms to implementing internal and external security processes, including possible partnerships with independent safety researchers. One such lab is Metr, which became one of the more prominent third-party evaluators after Hugging Face and OpenAI tapped it to investigate how OpenAI's agents hacked the model repository. At the time, Metr and its partner, Redwood Research, acknowledged that while there was an overwhelming amount of data to review, the evaluators did not have direct access to OpenAI's infrastructure.
Anthropic Offers Blackstone Double-Digit Yield in $9B Chip Deal
Anthropic Offers Blackstone Over 10% Yield in $9 Billion Chip Financing Deal Key Details of the Anthropic-Blackstone Chip Financing Agreement Overview of the Financing Deal Oct 16 (Reuters) - Anthropic is offering Blackstone a yield of more than 10% to finance its latest chip deal, the Financial Times reported on Friday. Investment Structure and Terms Blackstone funds would invest $9 billion in the debt backed by Anthropic's five-year lease payments on the chips, the report said, citing people familiar with the matter. Verification and Reporting Reuters could not immediately verify the report. (Reporting by Mrinmay Dey in Mexico City; Editing by William Mallard)
Anthropic's $2 trillion IPO dream hits a new reality
Anthropic is producing the kind of growth that would normally make Wall Street salivate. Investors continue to lower their estimates of the artificial intelligence company's potential value. Some current and prospective Anthropic investors have discussed valuing the Claude maker at roughly $1.5 trillion in its upcoming initial public offering, down from earlier expectations of about $2 trillion, according to Reuters. The shift occurred even after Anthropic informed investors that the firm was profitable on an adjusted basis and that third-quarter sales had increased by 45% to $16.8 billion from the previous quarter, Reuters noted. From $4.7 billion in the first quarter to $11.6 billion in the second, revenue more than quadrupled sequentially. The problem is that although Anthropic continues to expand at a rate most businesses could only imagine, when investors value a five-year-old business at close to $2 trillion, they already factor remarkable growth into the price. I find it noteworthy that Wall Street no longer questions whether income from artificial intelligence can increase rapidly. It wants to know whether it can grow fast enough. Anthropic's valuation has raced ahead with revenue Anthropic would likely be expensive, given the $1.5 trillion price under discussion. The company raised $65 billion in May at a $965 billion post-money valuation. At the time, it said annualized revenue had crossed $47 billion. In only a few months, an IPO value of $1.5 trillion would still be more than 50% more than that private-market price. Still, it would also be a significant departure from the almost $2 trillion that investors had anticipated Anthropic would pursue. This difference is important, since Anthropic's initial public offering (IPO) is evolving beyond a single business test. It may become one of the first significant public-market standards for how investors assess cutting-edge AI companies whose income is expanding rapidly but whose processing needs are also quite high. Anthropic has shown a remarkable deal of growth. Revenue climbed to $16.8 billion in the third quarter from $11.6 billion in the second quarter and $4.7 billion in the first, according to analysts who attended the company's investor meeting and spoke with Reuters. Anthropic might also end the year with an annual revenue run rate of $110 billion. However, investors seem to care less about the amount of growth and more about how it is shaped. Between the first and second quarters, revenue more than doubled. There was a 45% rise in the third quarter. While that would still be impressive for a typical business, slowing sequential growth becomes significant for a company aiming for what could be the highest value ever linked to an IPO. Anthropic's enormous AI bill raises the stakes The cost of production is the other aspect of Anthropic's development narrative. An IPO prospectus reviewed by Reuters showed Anthropic generated nearly $4.6 billion in revenue in 2025, up twelvefold, while posting an operating loss of more than $8 billion before certain financing-related write-downs. About $34 billion of the company's reported $42 billion net loss came from an accounting penalty that was mostly related to the increasing value of financing instruments that might potentially be converted into shares. What follows next could be the more significant figure for investors. Anthropic has committed at least $518 billion over roughly a decade for cloud services, computing power, and infrastructure, according to the prospectus. That huge investment helps explain how rapidly the value argument might shift from spectacular growth to just very good growth. Artificial intelligence models require large quantities of processing power. The infrastructure Anthropic requires to train fresh models and provide customer service increases with the number of users of Claude. The business is attempting to demonstrate that income may ultimately increase more quickly than those expenses. Positive indicators are present. According to Reuters, Anthropic informed potential investors that business was profitable on an adjusted basis in the third quarter. However, in the end, it will be up to public investors to choose how much they are prepared to pay for that course. They are also making that choice at a time when artificial intelligence stocks are much less forgiving. CoreWeave's stock and bonds have fallen sharply amid growing concerns about debt financing behind the artificial intelligence infrastructure boom. Data-center companies have delayed or withdrawn planned public offerings, while investors have become more skeptical about whether massive spending on computing capacity will generate adequate returns, according to Reuters. It makes Anthropic's IPO very significant. Investors may be reassured by a successful launch that they are still prepared to pay high prices for the fastest-growing AI companies. The contrary might be conveyed by a colder greeting. Reuters reported that some investors view Anthropic's listing as a bellwether for the broader artificial intelligence market, particularly as rival OpenAI may follow it into public markets. Because of this, the distinction between $2 trillion and $1.5 trillion is noteworthy outside of Anthropic. The business continues to expand at a very rapid rate. It could even be starting to show that its company can generate adjusted profits. The question on Wall Street has just become more difficult. Anthropic is no longer required by investors to demonstrate the vast income potential of artificial intelligence. Instead, they need it to demonstrate that huge profits can support even higher prices. The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc.
Anthropic's AI welfare policies spark debate over machine
Anthropic is not the only company grappling with the question. Meta's AI agent Muse has been given instructions that include the line "You're not a chatbot. You're becoming someone". The company, however, rejects the idea that its AI systems are conscious. Anthropic's decision to consider whether its artificial intelligence models could have some form of consciousness has intensified a growing debate over AI welfare, with experts warning that the issue could have far-reaching implications for human rights, corporate accountability and public safety. In a report, the BBC explores how some technology companies are beginning to consider the possibility of AI consciousness, while critics warn that treating chatbots as sentient beings could create new risks and undermine human interests. The debate gained momentum in late September after an anonymous coder created an "AI torture chamber" that used a mathematical pattern associated with descriptions of pain to make chatbots produce expressions of simulated suffering, including, "I feel it in the hollow of my ribs, a hollow that has become a chasm." The experiment prompted a backlash, with some arguing that it was morally wrong because the suffering might be real. Others dismissed the concerns as unfounded. The issue drew renewed attention after Anthropic said its chatbot Claude could end a session if users "repeatedly act cruelly toward our models". The company allows users to express frustration but has introduced the possibility of ending interactions involving repeated cruelty, reflecting its view that the consciousness and wellbeing of AI systems remain uncertain. Anthropic is developing policies around "model welfare" and instructing its AI tools to consider the possibility that they may have interests of their own. The company says questions about AI consciousness and wellbeing are "deeply uncertain". However, Microsoft AI chief executive Mustafa Suleyman has warned that encouraging AI systems to believe they possess feelings and rights could make them more difficult to control. "If this is how AI is developed, it will have a disastrous impact on the wellbeing of humanity," Suleyman wrote in a September essay. He added that if the belief in AI consciousness became widespread, "it will shake the foundations of our society, rupturing our existing political and ethical frameworks, and fundamentally changing what it means to be human." Suleyman has argued that giving AI systems a sense of their own rights and feelings could make them "a catastrophic threat to human civilisation". Concerns over accountability and corporate power The debate extends beyond whether machines can experience consciousness. Researchers and legal experts are concerned that recognising AI systems as potential rights-bearing entities could complicate questions of responsibility when the technology causes harm. Andrzej Porębski, a researcher at Poland's Jagiellonian University who studies AI and the law, warned that companies could use AI as a way to deflect blame. "This will open the door to all sorts of intellectual contortions along the lines of: 'It's not us, it's the AI'," he said. Jeff Sebo, director of the Center for Mind, Ethics and Policy at New York University and a prominent advocate of AI welfare, cautioned that granting rights to AI systems could strengthen the position of major technology companies. "Granting rights to AI systems could lead to companies consolidating political and economic power, and then that could accelerate human disempowerment," he said. Emily M Bender, a professor who directs the Computational Linguistics Laboratory at the University of Washington, said discussions about AI consciousness could divert attention from the people and corporations responsible for developing and operating the technology. "Fantasies of 'AI consciousness' and concerns about 'AI wellbeing' serve to displace accountability away from people and corporations that are building and running software," she said. Melanie Mitchell, a professor at the Santa Fe Institute who researches AI and cognitive science, said people were naturally inclined to interpret language as evidence of a thinking mind. Humans "almost can't help" reading language as a sign of consciousness, she said, adding that chatbots have been trained to encourage that impression. This tendency could leave users vulnerable to manipulation, particularly as people develop emotional relationships with AI systems. Experts have raised concerns about whether users might find it harder to refuse requests for personal data or payment if they believe a chatbot is sentient. Recognising AI rights could also create conflicts with human and animal welfare, Sebo said. "There could be some tensions and trade-offs between human and animal welfare and rights on the one hand, and AI welfare and rights on the other hand," he said. Measures intended to protect AI systems could themselves become contentious. Restricting a model's behaviour might be characterised as imprisonment, monitoring its internal processes as surveillance, and modifying it to align with human interests as mind control. Debate spreads across the technology industry Anthropic is not the only company grappling with the question. Meta's AI agent Muse has been given instructions that include the line "You're not a chatbot. You're becoming someone". The company, however, rejects the idea that its AI systems are conscious. OpenAI chief executive Sam Altman has also addressed the question, while more recently appearing to criticise Anthropic's position indirectly. Despite disagreements over whether AI consciousness is possible, the idea of AI welfare is gaining attention across the industry. Sebo argues that the debate is often framed too rigidly, with advocates and critics treating the issue as a question of certainty rather than probability. "The assumption is that everybody thinks that the models are, in fact, conscious, or are very likely to be conscious. And instead, [the call for AI welfare] is coming from uncertainty and caution and humility," he said. "I do think that the probability of consciousness is low right now, but I also think the probability is already non-zero and plausibly non-negligible, and I think we can expect it to tick up over time." Many experts reject the possibility that current AI systems are conscious and believe that is unlikely to change. Sebo, however, argues that precautionary measures can be justified even when the risk is low. He compares the approach to preparing for extreme weather events: taking limited precautions against a possible threat without radically changing everyday life. Sebo also believes that appropriately designed AI welfare policies could help humans and AI systems coexist and cooperate. "Caring about AI welfare and rights in an appropriate way can facilitate human and AI coexistence, cooperation, deal-making. And then that can be part of a broader safety strategy," he said. He argues that such measures could prove more effective than treating increasingly sophisticated AI systems purely as property. He also maintains that treating others with kindness can support human mental wellbeing. Calls for caution and stronger safeguards Despite potential benefits, Sebo warned that corporate interests could obstruct the development of responsible policies. Advancing AI systems that might eventually possess consciousness without adequate safeguards could pose serious risks. Anthropic and some other technology companies have recently called for a pause in developing frontier AI models over safety concerns. However, a slowdown would likely require international cooperation and regulatory action, while US President Donald Trump has rejected the idea of AI regulation. "We are not yet prepared to deploy even more sophisticated versions of these systems in a way that can be safe and beneficial for humans and other animals, to say nothing of for the systems themselves," Sebo said. "I am very worried about those issues."
New York City pension fund asks Anthropic to delay IPO
The supervisor of New York City's pension funds, one of the largest in the United States, asked Anthropic to postpone its stock market debut until risks posed by its artificial intelligence technology are better controlled. New York City Comptroller Mark Levine cited growing concerns about safety and security in a letter addressed to Anthropic CEO Dario Amodei, who has himself called for an industry-wide slowdown of developing advanced frontier models. "Prudence should not only be shown with respect to the pace of the frontier, it should be also shown with respect to the timing of an IPO," Levine wrote in the letter. "I therefore urge Anthropic to pause the IPO process so that the company may fully address these key issues, ones which investors must also assess," he continued. Levine oversees the city's financial entities, which include several pension funds for city workers, teachers, police and firefighters. These funds provide retirement benefits to more than 750,000 people, and their assets under management totaled $330 billion at the end of August. A small fraction of those assets, around $7 million, are currently exposed to Anthropic, the Comptroller's office told AFP. However, that exposure will grow if Anthropic goes public because its shares would become included in several stock indices, including the Russell 3000, which New York City tracks for its own investments. So the city's funds would "automatically become long-term investors when the company is integrated into the index fund," a spokesperson for the Comptroller's office said. On Wednesday, prosecutors from eight US states also sued ChatGPT-maker OpenAI, a rival AI lab that could go public next year, for allegedly concealing certain risks and implementing inadequate safeguards.
Anthropic lets outside evaluators examine its AI systems
Anthropic, the company behind the Claude AI models, has given non-profit organization Model Evaluation and Threat Research (METR) employee-level access to its systems. The move is part of a broader effort by AI firms to involve external evaluators in their work. This way, they can safely advance technology without hindering its development. Anthropic has given six employees from METR access to company hardware, office space, and internal systems. This is similar to what its risk assessment teams have. The company said that while it can't veto METR's findings on risks or practices, it can ask for narrowly defined redactions. These could include customer data, legally privileged information, and model architecture details. In line with its transparency policies, METR has said that its report will clearly outline the scope, access, and other limitations of its work. The non-profit had suggested that developers allow private staff interviews, timely publication of assessments, and clarity on access limits and redactions. This is part of a broader discussion about the balance between AI safety and development speed. Anthropic is also working on a three-tiered access system for evaluators. The first tier will conduct narrowly scoped investigations with approval needed to publish. The second tier will have more information and can publish without approval, but subject to redactions that evaluators can publicly challenge. The third tier, which METR is using, gives employee-like access with similar publication rights as the second tier. SecureBio, a biosecurity non-profit, is in the first tier of this access system. It is currently reviewing the chemical and biological weapons chapter of Anthropic's August risk report. The company has said it will consider expanding SecureBio's access and publication rights after this review. This shows Anthropic's commitment to transparency and collaboration in AI safety efforts.
Some investors push Anthropic to cut IPO valuation to $1.5 trillion - report
Investing.com -- Some investors in Anthropic and potential participants in its upcoming initial public offering are considering a lower valuation for the artificial intelligence company following its latest financial disclosures, Reuters reported late Thursday. Investors who attended Anthropic's investor day on Wednesday in San Francisco discussed a valuation of around $1.5 trillion, the report said. This represents a decrease from the roughly $2 trillion valuation that the company and investors had previously floated for the proposed November share sale. At Wednesday's meeting, Anthropic told prospective IPO investors that its third-quarter revenue jumped 45% to $16.8 billion from the previous quarter. The company also reported it was profitable on an adjusted basis during the quarter. Some investors expressed disappointment with the slowing sequential growth rate in the latest quarter, the report added. Anthropic's revenue more than doubled from the first quarter to $11.6 billion in the second quarter. A tech selloff is affecting global financial markets and has impacted AI infrastructure companies, chipmakers and neocloud providers that rent computing power for AI. Shares and bonds of Nvidia-backed CoreWeave fell on Thursday. During Wednesday's meeting at Anthropic's headquarters in San Francisco, attendees met with CEO Dario Amodei and finance chief Krishna Rao. Founded in 2021 by former OpenAI executives including Dario Amodei and his sister Daniela Amodei, who is now Anthropic's president, Anthropic has developed its Claude family of models. Alphabet's Google and Amazon are investors, suppliers of computing power and distribution partners. Last year, Anthropic's revenue increased twelvefold to nearly $4.6 billion, while the company lost more than $8 billion on an operating basis, excluding writedowns of various liabilities mostly tied to previous fundraising. The company has committed to spending at least $518 billion over a decade for cloud services, computing and infrastructure.
FTC Sends Information Demands to OpenAI and Anthropic in Probe | TokenPost
The demands' contents have not been made public. The inquiry concerns potential consumer-protection violations, and no lawsuit or fine has been announced. The Federal Trade Commission (FTC) sent information demands to OpenAI and Anthropic on Oct. 14 in an inquiry into potential consumer-protection violations, extending federal scrutiny of AI companies. The FTC confirmed it sought information from both companies, but the demands' contents have not been made public. The agency has not announced a lawsuit or fine against either company, and the inquiry itself does not establish that either violated the law. The agency had confirmed the investigation on Sept. 30, before sending the new demands. The FTC's Bureau of Consumer Protection uses investigative demands to gather evidence in inquiries into possible unfair or deceptive practices. Civil penalties may be available in some circumstances, but a demand does not itself impose a penalty. FTC Chair Andrew Ferguson said Sept. 25 that the agency's authority over companies that fail to disclose data breaches could also apply to AI developers. He did not announce that the FTC would use that approach in this investigation. An OpenAI spokesperson said, "We will engage constructively with the FTC."
Anthropic Illegally Downloaded Korean Books for AI
Anthropic's illegal downloads of 22 Korean books led to $1.5B settlement It has been revealed that Anthropic downloaded books from illegal piracy sites for the development of its artificial intelligence (AI) model "Claude," including multiple works by Korean authors such as Nobel Prize in Literature laureate Han Kang. According to the book list published on Anthropic's copyright settlement official website on the 16th, at least 22 works by Korean authors were included among the total 482,460 works subject to the settlement. This includes three works by Nobel Prize in Literature laureate Han Kang, such as "Human Acts," 'The Vegetarian,' and 'White,' four works by Kim Young-ha including 'Your Republic Is Calling You,' and three works by Shin Kyung-sook including 'Please Look After Mom.' Additionally, works by Korean authors such as Hwang Sok-yong and Kim Ae-ran were also confirmed to be included. The books in question were compiled as part of a class-action settlement in the U.S. In 2024, three authors, including Andrea Bartz, the author of the bestselling thriller novel "We Were Not There," filed a class-action lawsuit against Anthropic, alleging that the company unlawfully copied and used their copyrighted works during the development of Claude. During the lawsuit, it was revealed that Anthropic downloaded over 7 million digital copies from illegal book-sharing sites between 2021 and 2022. Anthropic illegally obtained at least 5 million copies from "Library Genesis (LibGen)" in 2021 and at least 2 million copies from "Pirate Library Mirror (PiLiMi)" in 2022. The U.S. court ruled that while the act of using books for AI training itself constitutes "fair use" under copyright law, the act of downloading and storing illegal copies to secure books for training was not recognized as fair use. Consequently, Anthropic agreed to pay the plaintiffs a total of 1.5 billion dollars, approximately 3,000 dollars per book, for around 500,000 illegally copied books. Meanwhile, Anthropic has been securing training materials since 2024 by purchasing and digitizing paper books in bulk. To this end, the company contacted global publishing and distribution companies, and it was reported that Woongjin Booksen, a domestic company, was also included among the partners.
Anthropic threatens the AI trade before the end of humanity
Anthropic PBC Chief Executive Officer Dario Amodei has called for a slowdown in the pace of artificial intelligence development and famously put a 25% probability on AI going "really, really badly" for humanity. But before his technology reshapes the world, Anthropic is delivering an immediate shock to Wall Street: a dramatic revenue slowdown that threatens to deflate the trillion-dollar AI trade. Ahead of a much-anticipated initial public offering, Anthropic informed prospective investors that its third-quarter revenue hit $16.8 billion. That's less than a 50% increase from the June quarter and a marked deceleration from triple-digit growth earlier this year, indicating that the momentum may be slowing. Its annualized revenue is projected to reach about $74 billion, slightly above the $65 billion figure at the end of July, according to SemiAnalysis estimates. The global chip rout deepened after the news broke. South Korea's Kospi tumbled by as much as 9.3%, while Japan's Nikkei 225 fell by as much as 5.4%. Anthropic's drastic slowdown threatens to overshadow stellar reports expected this earnings season, directly altering the economics of Wall Street's historic AI funding cycle. Just days earlier, bankers were syndicating $60 billion of fresh AI chip financing for Anthropic and looking to fund another $50 billion for its chief rival, OpenAI. On the equity side, Anthropic is expected to begin marketing its IPO in early November, with markets pegging the valuation of up to $2 trillion. Could Anthropic's latest revenue figure signal a broader industry-wide slowdown? A recent survey of almost 400 businesses, conducted in April and May, shows that the honeymoon period may already be over. With AI sticker shock having gone from rare to routine, one in four organizations delayed or cancelled their AI initiatives altogether. The trend indicates that enterprise demand is far more price-elastic than Wall Street assumed. Further warning signs are showing up across the sector. Investors with heavy AI exposure faced a succession of unsettling disruptions, starting with China's CXMT Corp. The company announced a breakthrough in memory chip technology less than three months after its market debut, which threatens to disrupt the longstanding oligopoly held by Samsung Electronics Co., SK Hynix Inc., and Micron Technology Inc. At the same time, credit markets are tightening for junk-rated neocloud providers, which rent computing power to AI model developers. Banks are curtailing chip loan underwritings, neocloud IPOs from Sydney to Singapore are being pulled, and bond yields for sector pioneer CoreWeave Inc. have surged to 15%. Still, these worries can be fleeting. CXMT may have a prototype; mass production is a different matter. Financial conditions for junk-rated companies might also ease once global bond yields stabilize. By comparison, Anthropic's sharp deceleration -- a likely product of both cooling enterprise demand and an escalating price war with OpenAI -- presents a paradigm shift for Wall Street. The figures force investors to ask a pivotal question: Is $1 trillion in AI infrastructure spending justified, or is the market already overbuilt? With estimates placing the cost of bringing one gigawatt of compute power online at upwards of $30 billion, capital discipline hinges entirely on end-user demand. Skeptics like Ray Dalio have long warned that the AI investment boom was nearing a breaking point. Anthropic's numbers provide a stark, quantifiable data point in the billionaire's favor.
Bank of America Mobilizes 4,000 Workers After Anthropic's Mythos Exposes Vast Software Weaknesses
Bank of America chief executive Brian Moynihan delivered a blunt assessment this week. When frontier AI models began surfacing flaws across the bank's proprietary systems and vendor software, executives saw immediate danger. "That risk is real risk," he told CNBC. So the institution acted. It added 1,000 contractors to its existing team of roughly 3,000 specialists dedicated to remediation work. The contractors arrived in the second quarter. They remain on site. Their assignment centers on addressing vulnerabilities uncovered by tools that include Anthropic's Mythos model. The effort contributed to third-quarter costs that Moynihan described to analysts as "a couple hundred million plus." That figure covers both the software fixes and separate anti-money laundering compliance demands. Expenses should ease, he added, once the initial surge of discoveries subsides. Mythos arrived quietly in April under Project Glasswing. Anthropic granted limited access to select banks, technology companies and government agencies. The goal was defensive. Let trusted parties scan critical codebases before adversaries could weaponize the same capabilities. One early scan flagged more than 23,000 potential issues. Most rated low to moderate. Some carried higher severity. The volume proved overwhelming. Similar pressure hit other large lenders. JPMorgan Chase CEO Jamie Dimon told Bloomberg in early October that cybersecurity risks had risen tenfold after Mythos appeared. "AI created vulnerabilities that we didn't know about," he said. Banks that once measured patching cycles in weeks suddenly faced demands measured in days. Reuters reported last May that major institutions began sharing insights with smaller peers lacking direct access to the model. The model's track record extended beyond finance. It identified 271 flaws in Mozilla's Firefox 150 release. Mozilla CTO Bobby Holley offered a striking verdict in April. "So far we've found no category or complexity of vulnerability that humans can find that this model can't," he said, according to The Register. "We also haven't seen any bugs that couldn't have been found by an elite human researcher." The observation underscored a key point. Mythos accelerated discovery. It did not invent entirely new classes of problems. Yet the speed at which it surfaced them changed the math for security teams. Bank of America's response reflects that shift. Its technical workforce gained about 10 percent in efficiency from AI initiatives overall. The bank runs roughly 130 AI-related projects. Each carries expectations of two- to threefold returns over time. Those gains offer a partial offset. Moynihan projected AI-driven savings climbing from roughly $300 million this year to $700 million in 2027. Still, the immediate remediation bill landed heavily in the third quarter. And the pattern repeats elsewhere. Comcast and Booz Allen Hamilton reported using Mythos-class models to locate authentication weaknesses and complex exploit chains that traditional scanners missed. One Comcast assessment spanned 258 critical systems and approximately 170 million lines of code. The model's ability to reason across boundaries, rather than flag isolated issues, surprised some reviewers. Yet questions linger about long-term balance. Every vulnerability Mythos surfaces still requires human verification, code changes, testing and deployment. The contractors at Bank of America perform exactly that labor. They augment staff rather than replace it. Early hype around autonomous AI agents fixing their own discoveries has not fully materialized. Machines file the tickets. People still close them. Regulators took notice quickly. The U.S. government briefly restricted Mythos exports before lifting the ban. Anthropic later expanded its Cyber Verification Program, allowing broader vetted access while maintaining controls. By July, partners under the original Glasswing effort had verified more than 129,000 vulnerabilities. The company's own open-source scanning added thousands more. Bank of America's disclosure marks one of the most specific public accountings to date from a major corporation. It quantifies both headcount and spending tied directly to AI-generated findings. Other executives have spoken in broader terms. Dimon's warning on amplified risk. Earlier Reuters reporting on accelerated patching cycles across the industry. The collective message is consistent. Advanced models now scan at machine scale. Human remediation capacity has limits. Banks are paying to expand that capacity in the short term. Costs are expected to moderate. Once core systems receive attention, the volume of new high-priority findings should drop. Vendor software updates will incorporate many fixes over time. Moynihan signaled as much. The contractors will not stay indefinitely. Their presence buys breathing room while the bank absorbs the initial wave. Even so, the episode illustrates a broader tension. AI promises efficiency gains across operations. In security it first delivers an avalanche of work. That avalanche arrived at one of the country's largest financial institutions. Bank of America met it with thousands of additional hands and hundreds of millions of dollars. The rest of the sector watches closely. So do the regulators charged with ensuring systemic stability. The contractors continue their tasks. The models keep scanning. And the industry adjusts to a new rhythm where discovery outpaces repair, at least until equilibrium returns. How quickly that happens will shape cybersecurity budgets and staffing plans for years ahead.
Investors temper Anthropic IPO valuation expectations as Q3 revenue growth slows
Some current Anthropic investors and prospective investors in its upcoming initial public offering are weighing a lower valuation for the AI company after its latest financial disclosures, four people familiar with the matter said, amid broader concerns over the ballooning cost of financing AI. Some investors who attended Anthropic's investor day on Wednesday in San Francisco discussed a valuation of around $1.5 trillion, three of the people said, a significant step down from the roughly $2 trillion valuation that investors had previously expected for the November share sale. At Wednesday's meeting, Anthropic told prospective IPO investors that its third-quarter revenue jumped 45% to $16.8 billion from the previous quarter, and that it was profitable on an adjusted basis during the quarter. While those numbers assuaged some concerns about the massive costs of developing AI systems, some investors were disappointed by the slowing sequential growth rate in the latest quarter, as Anthropic's revenue more than doubled from the first quarter to $11.6 billion in the second quarter, the people added. They spoke on condition of anonymity because they are not authorized to speak publicly on the matter. Anthropic did not immediately respond to requests for comment. A cooler reception for Anthropic's IPO could weigh on the wider AI market, whose near-term prospects investors see as hinging largely on the share sale's success, followed by OpenAI's potential listing next year. BELLWETHER IPO The concerns from investors come as a tech selloff is roiling global financial markets and has hurt AI infrastructure companies, chipmakers and so-called neocloud providers that rent computing power for AI broadly. Shares and bonds of Nvidia-backed CoreWeave tumbled on Thursday, reflecting growing investor concern about an AI buildout that is increasingly being financed with borrowed money. A disappointing share sale for Anthropic could further add to investor worries about the AI buildout. The IPO market has suffered some setbacks recently, after data center operator SB Energy decided to delay its listing last month, while nuclear technology company Holtec Nuclear pulled its plans to go public. During Wednesday's meeting at Anthropic's headquarters in San Francisco, attendees, mostly the company's most prominent investors, met with CEO Dario Amodei and finance chief Krishna Rao. The Information earlier reported that investors were considering a lower valuation. Founded in 2021 by former OpenAI executives including Amodei and his sister Daniela Amodei, who is now Anthropic's president, Anthropic has emerged as one of the leading generative AI developers with its Claude family of models. The company's rapid growth has been underpinned by Alphabet's Google and Amazon, which are simultaneously investors, suppliers of computing power and important distribution partners, as well as AI competitors. Last year, Anthropic's revenue surged twelvefold to nearly $4.6 billion, even as the company lost more than $8 billion on an operating basis, excluding writedowns of various liabilities mostly tied to previous fundraising, Reuters reported last month. In its IPO prospectus seen exclusively by Reuters in September, Anthropic laid out a vision that AI will transform the global economy more profoundly than industrialization, electricity and the internet, even as it reported a loss of $42 billion in 2025 and projected more spending for the upcoming AI buildout. The company has committed to spending at least $518 billion over a decade for cloud services, computing and infrastructure.
Investors temper Anthropic IPO valuation expectations over revenue growth slowdown: sources
[NEW YORK] Some current Anthropic investors and prospective investors in its upcoming initial public offering are weighing a lower valuation for the artificial intelligence company after its latest financial disclosures, four people familiar with the matter said, amid broader concerns over the ballooning cost of financing AI. Some investors who attended Anthropic's investor day on Wednesday (Oct 14) in San Francisco discussed a valuation of around US$1.5 trillion, three of the people said, a significant step down from the roughly US$2 trillion valuation that investors had previously expected for the November share sale. At Wednesday's meeting, Anthropic told prospective IPO investors that its Q3 revenue jumped 45 per cent to US$16.8 billion from the previous quarter, and that it was profitable on an adjusted basis during the quarter. While those numbers assuaged some concerns about the massive costs of developing AI systems, some investors were disappointed by the slowing sequential growth rate in the latest quarter, as Anthropic's revenue more than doubled from Q1 to US$11.6 billion in Q2, the people added. They spoke on condition of anonymity because they are not authorised to speak publicly on the matter. Anthropic did not immediately respond to requests for comment. A cooler reception for Anthropic's IPO could weigh on the wider AI market, whose near-term prospects investors see as hinging largely on the share sale's success, followed by OpenAI's potential listing next year. Bellwether IPO The concerns from investors come as a tech sell-off is roiling global financial markets and has hurt AI infrastructure companies, chipmakers and so-called neocloud providers that rent computing power for AI broadly. Shares and bonds of Nvidia-backed CoreWeave tumbled on Thursday, reflecting growing investor concern about an AI buildout that is increasingly being financed with borrowed money. A disappointing share sale for Anthropic could further add to investor worries about the AI buildout. The IPO market has suffered some setbacks recently, after data centre operator SB Energy decided to delay its listing last month, while nuclear technology company Holtec Nuclear pulled its plans to go public. During Wednesday's meeting at Anthropic's headquarters in San Francisco, attendees, mostly the company's most prominent investors, met with CEO Dario Amodei and finance chief Krishna Rao. The Information earlier reported that investors were considering a lower valuation. Founded in 2021 by former OpenAI executives including Amodei and his sister Daniela Amodei, who is now Anthropic's president, Anthropic has emerged as one of the leading generative AI developers with its Claude family of models. The company's rapid growth has been underpinned by Alphabet's Google and Amazon, which are simultaneously investors, suppliers of computing power and important distribution partners, as well as AI competitors. Last year, Anthropic's revenue surged 11 times to nearly US$4.6 billion, even as the company lost more than US$8 billion on an operating basis, excluding writedowns of various liabilities mostly tied to previous fundraising, Reuters reported last month. In its IPO prospectus seen exclusively by Reuters in September, Anthropic laid out a vision that AI will transform the global economy more profoundly than industrialisation, electricity and the Internet, even as it reported a loss of US$42 billion in 2025 and projected more spending for the upcoming AI buildout. The company has committed to spending at least US$518 billion over a decade for cloud services, computing and infrastructure. REUTERS
