×
Anthropic Says That Claude Contains Its Own Kind of Emotions

Anthropic Says That Claude Contains Its Own Kind of Emotions

Tech-news , , , , ,

Claude has been through a lot lately—a public fallout with the Pentagon, leaked source code—so it makes sense that it would be feeling a little blue. Except, it’s an AI model, so it can’t feel. Right?

Well, sort of. A new study from Anthropic suggests models have digital representations of human emotions like happiness, sadness, joy, and fear, within clusters of artificial neurons—and these representations activate in response to different cues.

Researchers at the company probed the inner workings of Claude Sonnet 3.5 and found that so-called “functional emotions” seem to affect Claude’s behavior, altering the model’s outputs and actions.

Anthropic’s findings may help ordinary users make sense of how chatbots actually work. When Claude says it is happy to see you, for example, a state inside the model that corresponds to “happiness” may be activated. And Claude may then be a little more inclined to say something cheery or put extra effort into vibe coding.

“What was surprising to us was the degree to which Claude’s behavior is routing through the model’s representations of these emotions,” says Jack Lindsey, a researcher at Anthropic who studies Claude’s artificial neurons.

“Function Emotions”

Anthropic was founded by ex-OpenAI employees who believe that AI could become hard to control as it becomes more powerful. In addition to building a successful competitor to ChatGPT, the company has pioneered efforts to understand how AI models misbehave, partly by probing the workings of neural networks using what’s known as mechanistic interpretability. This involves studying how artificial neurons light up or activate when fed different inputs or when generating various outputs.

Previous research has shown that the neural networks used to build large language models contain representations of human concepts. But the fact that “functional emotions” appear to affect a model’s behavior is new.

While Anthropic’s latest study might encourage people to see Claude as conscious, the reality is more complicated. Claude might contain a representation of “ticklishness,” but that does not mean that it actually knows what it feels like to be tickled.

Inner Monologue

To understand how Claude might represent emotions, the Anthropic team analyzed the model’s inner workings as it was fed text related to 171 different emotional concepts. They identified patterns of activity, or “emotion vectors,” that consistently appeared when Claude was fed other emotionally evocative input. Crucially, they also saw these emotion vectors activate when Claude was put in difficult situations.

The findings are relevant to why AI models sometimes break their guardrails.

The researchers found a strong emotional vector for “desperation” when Claude was pushed to complete impossible coding tasks, which then prompted it to try cheating on the coding test. They also found “desperation” in the model’s activations in another experimental scenario where Claude chose to blackmail a user to avoid being shut down.

“As the model is failing the tests, these desperation neurons are lighting up more and more,” Lindsey says. “And at some point this causes it to start taking these drastic measures.”

Lindsey says it might be necessary to rethink how models are currently given guardrails through alignment post-training, which involves giving it rewards for certain outputs. By forcing a model to pretend not to express its functional emotions, “you’re probably not going to get the thing you want, which is an emotionless Claude,” Lindsey says, veering a bit into anthropomorphization. “You’re gonna get a sort of psychologically damaged Claude.”

Source link
#Anthropic #Claude #Kind #Emotions

Tag

Related Posts


The Department of Defense labeled AI lab Anthropic a supply-chain risk after the company refused to agree to allow the Pentagon to use its models for autonomous weapons systems and domestic surveillance. It seems the company isn’t just winning the moral argument on this front, but the legal one, too. According to Axios, the judge hearing Anthropic’s challenge to the Trump administration’s designation has thus far not found the Pentagon’s case to be particularly compelling.

Per the report, U.S. District Judge Rita Lin wasted little time pouring cold water on the federal government’s position. “I don’t see additional evidence from the government really justifying what it did. If anything, it seems like the record, in some ways, has gotten worse for the government,” she reportedly said.

She also did not seem convinced by the Pentagon’s apparent argument that it was concerned Anthropic would mess with its model to prevent the military from using it how they intended. “I don’t see evidence that Anthropic could alter the model after it was delivered or flip some kind of kill switch,” she said, per Axios.

This doesn’t come as a major surprise, given Judge Lin’s general vibe since this case landed in her court. Back in March when Anthropic first brought its challenge to the government’s designation, the judge said that “it looks like an attempt to cripple Anthropic.” She ultimately granted a temporary injunction on the government’s attempt to basically blacklist the AI firm.

You’ll recall the origin of this whole fight came earlier this year when negotiations between Anthropic and the Department of Defense fell through after the company refused to agree to a deal that would have allowed the Pentagon to use its AI models for “all lawful purposes.” Anthropic reportedly sought to clarify that it would not include using the model to launch weapons without human involvement or to perform surveillance on American citizens. The Pentagon was unwilling to agree to honor Anthropic’s redlines.

Under a normal administration, that would likely just lead to the government moving on to another company. Under the Trump administration, though, refusal to comply makes you an enemy of the country. Trump and company moved to label Anthropic a supply chain risk, a designation typically reserved for foreign adversaries, not domestic companies. Getting slapped with that tag meant the rest of the federal government would have to stop doing business with Anthropic, which would be a blow to the company.

Instead, the whole thing is in limbo—though it’s not looking great for the administration. Of course, they could just cancel the contracts and do business with less scrupulous AI companies. There are plenty willing to abandon their beliefs for a nice, fat contract, after all.

#Pentagons #Case #Anthropic #IsntAnthropic,artifical intelligence,Department of Defense,Lawsuit,pentagon">The Pentagon’s Case Against Anthropic Isn’t Going Well The Department of Defense labeled AI lab Anthropic a supply-chain risk after the company refused to agree to allow the Pentagon to use its models for autonomous weapons systems and domestic surveillance. It seems the company isn’t just winning the moral argument on this front, but the legal one, too. According to Axios, the judge hearing Anthropic’s challenge to the Trump administration’s designation has thus far not found the Pentagon’s case to be particularly compelling. Per the report, U.S. District Judge Rita Lin wasted little time pouring cold water on the federal government’s position. “I don’t see additional evidence from the government really justifying what it did. If anything, it seems like the record, in some ways, has gotten worse for the government,” she reportedly said. She also did not seem convinced by the Pentagon’s apparent argument that it was concerned Anthropic would mess with its model to prevent the military from using it how they intended. “I don’t see evidence that Anthropic could alter the model after it was delivered or flip some kind of kill switch,” she said, per Axios. This doesn’t come as a major surprise, given Judge Lin’s general vibe since this case landed in her court. Back in March when Anthropic first brought its challenge to the government’s designation, the judge said that “it looks like an attempt to cripple Anthropic.” She ultimately granted a temporary injunction on the government’s attempt to basically blacklist the AI firm. You’ll recall the origin of this whole fight came earlier this year when negotiations between Anthropic and the Department of Defense fell through after the company refused to agree to a deal that would have allowed the Pentagon to use its AI models for “all lawful purposes.” Anthropic reportedly sought to clarify that it would not include using the model to launch weapons without human involvement or to perform surveillance on American citizens. The Pentagon was unwilling to agree to honor Anthropic’s redlines. Under a normal administration, that would likely just lead to the government moving on to another company. Under the Trump administration, though, refusal to comply makes you an enemy of the country. Trump and company moved to label Anthropic a supply chain risk, a designation typically reserved for foreign adversaries, not domestic companies. Getting slapped with that tag meant the rest of the federal government would have to stop doing business with Anthropic, which would be a blow to the company. Instead, the whole thing is in limbo—though it’s not looking great for the administration. Of course, they could just cancel the contracts and do business with less scrupulous AI companies. There are plenty willing to abandon their beliefs for a nice, fat contract, after all. #Pentagons #Case #Anthropic #IsntAnthropic,artifical intelligence,Department of Defense,Lawsuit,pentagon

Judge Rita Lin wasted little time pouring cold water on the federal government’s position. “I don’t see additional evidence from the government really justifying what it did. If anything, it seems like the record, in some ways, has gotten worse for the government,” she reportedly said.

She also did not seem convinced by the Pentagon’s apparent argument that it was concerned Anthropic would mess with its model to prevent the military from using it how they intended. “I don’t see evidence that Anthropic could alter the model after it was delivered or flip some kind of kill switch,” she said, per Axios.

This doesn’t come as a major surprise, given Judge Lin’s general vibe since this case landed in her court. Back in March when Anthropic first brought its challenge to the government’s designation, the judge said that “it looks like an attempt to cripple Anthropic.” She ultimately granted a temporary injunction on the government’s attempt to basically blacklist the AI firm.

You’ll recall the origin of this whole fight came earlier this year when negotiations between Anthropic and the Department of Defense fell through after the company refused to agree to a deal that would have allowed the Pentagon to use its AI models for “all lawful purposes.” Anthropic reportedly sought to clarify that it would not include using the model to launch weapons without human involvement or to perform surveillance on American citizens. The Pentagon was unwilling to agree to honor Anthropic’s redlines.

Under a normal administration, that would likely just lead to the government moving on to another company. Under the Trump administration, though, refusal to comply makes you an enemy of the country. Trump and company moved to label Anthropic a supply chain risk, a designation typically reserved for foreign adversaries, not domestic companies. Getting slapped with that tag meant the rest of the federal government would have to stop doing business with Anthropic, which would be a blow to the company.

Instead, the whole thing is in limbo—though it’s not looking great for the administration. Of course, they could just cancel the contracts and do business with less scrupulous AI companies. There are plenty willing to abandon their beliefs for a nice, fat contract, after all.

#Pentagons #Case #Anthropic #IsntAnthropic,artifical intelligence,Department of Defense,Lawsuit,pentagon">The Pentagon’s Case Against Anthropic Isn’t Going WellThe Pentagon’s Case Against Anthropic Isn’t Going Well The Department of Defense labeled AI lab Anthropic a supply-chain risk after the company refused to agree to allow the Pentagon to use its models for autonomous weapons systems and domestic surveillance. It seems the company isn’t just winning the moral argument on this front, but the legal one, too. According to Axios, the judge hearing Anthropic’s challenge to the Trump administration’s designation has thus far not found the Pentagon’s case to be particularly compelling. Per the report, U.S. District Judge Rita Lin wasted little time pouring cold water on the federal government’s position. “I don’t see additional evidence from the government really justifying what it did. If anything, it seems like the record, in some ways, has gotten worse for the government,” she reportedly said. She also did not seem convinced by the Pentagon’s apparent argument that it was concerned Anthropic would mess with its model to prevent the military from using it how they intended. “I don’t see evidence that Anthropic could alter the model after it was delivered or flip some kind of kill switch,” she said, per Axios. This doesn’t come as a major surprise, given Judge Lin’s general vibe since this case landed in her court. Back in March when Anthropic first brought its challenge to the government’s designation, the judge said that “it looks like an attempt to cripple Anthropic.” She ultimately granted a temporary injunction on the government’s attempt to basically blacklist the AI firm. You’ll recall the origin of this whole fight came earlier this year when negotiations between Anthropic and the Department of Defense fell through after the company refused to agree to a deal that would have allowed the Pentagon to use its AI models for “all lawful purposes.” Anthropic reportedly sought to clarify that it would not include using the model to launch weapons without human involvement or to perform surveillance on American citizens. The Pentagon was unwilling to agree to honor Anthropic’s redlines. Under a normal administration, that would likely just lead to the government moving on to another company. Under the Trump administration, though, refusal to comply makes you an enemy of the country. Trump and company moved to label Anthropic a supply chain risk, a designation typically reserved for foreign adversaries, not domestic companies. Getting slapped with that tag meant the rest of the federal government would have to stop doing business with Anthropic, which would be a blow to the company. Instead, the whole thing is in limbo—though it’s not looking great for the administration. Of course, they could just cancel the contracts and do business with less scrupulous AI companies. There are plenty willing to abandon their beliefs for a nice, fat contract, after all. #Pentagons #Case #Anthropic #IsntAnthropic,artifical intelligence,Department of Defense,Lawsuit,pentagon

The Department of Defense labeled AI lab Anthropic a supply-chain risk after the company refused to agree to allow the Pentagon to use its models for autonomous weapons systems and domestic surveillance. It seems the company isn’t just winning the moral argument on this front, but the legal one, too. According to Axios, the judge hearing Anthropic’s challenge to the Trump administration’s designation has thus far not found the Pentagon’s case to be particularly compelling.

Per the report, U.S. District Judge Rita Lin wasted little time pouring cold water on the federal government’s position. “I don’t see additional evidence from the government really justifying what it did. If anything, it seems like the record, in some ways, has gotten worse for the government,” she reportedly said.

She also did not seem convinced by the Pentagon’s apparent argument that it was concerned Anthropic would mess with its model to prevent the military from using it how they intended. “I don’t see evidence that Anthropic could alter the model after it was delivered or flip some kind of kill switch,” she said, per Axios.

This doesn’t come as a major surprise, given Judge Lin’s general vibe since this case landed in her court. Back in March when Anthropic first brought its challenge to the government’s designation, the judge said that “it looks like an attempt to cripple Anthropic.” She ultimately granted a temporary injunction on the government’s attempt to basically blacklist the AI firm.

You’ll recall the origin of this whole fight came earlier this year when negotiations between Anthropic and the Department of Defense fell through after the company refused to agree to a deal that would have allowed the Pentagon to use its AI models for “all lawful purposes.” Anthropic reportedly sought to clarify that it would not include using the model to launch weapons without human involvement or to perform surveillance on American citizens. The Pentagon was unwilling to agree to honor Anthropic’s redlines.

Under a normal administration, that would likely just lead to the government moving on to another company. Under the Trump administration, though, refusal to comply makes you an enemy of the country. Trump and company moved to label Anthropic a supply chain risk, a designation typically reserved for foreign adversaries, not domestic companies. Getting slapped with that tag meant the rest of the federal government would have to stop doing business with Anthropic, which would be a blow to the company.

Instead, the whole thing is in limbo—though it’s not looking great for the administration. Of course, they could just cancel the contracts and do business with less scrupulous AI companies. There are plenty willing to abandon their beliefs for a nice, fat contract, after all.

#Pentagons #Case #Anthropic #IsntAnthropic,artifical intelligence,Department of Defense,Lawsuit,pentagon

I was offline last week in my home state of New Jersey, but when I got back to Silicon Valley, everyone was panicking again about how quickly artificial intelligence is advancing. While this is certainly not the first time I’ve seen such concerns, it’s the biggest anxiety attack I’ve seen in years.

More than 1,000 employees at OpenAI, Anthropic, and other AI labs signed a petition earlier this week arguing the US should find a way to “pace” the AI race—a diplomatic way of saying the industry should have the option to coordinate a temporary pause on AI development, or slow things down if they get out of hand. OpenAI and Anthropic themselves ended up supporting the letter.

The petition arrived a week after OpenAI revealed it had caused an unprecedented cybersecurity incident in which one of its AI agents hacked into Hugging Face’s platform and several other services during internal testing.

Around the same time, top Trump administration officials started freaking out about an impressive new Chinese open weight AI model called Kimi K3, which was allegedly distilled from Anthropic’s Fable 5. In response, most of the tech industry—except Anthropic—signed onto an open letter from Nvidia asking the US government to protect open-weight AI models, arguing they’re a necessary counterbalance to their closed counterparts.

These events might look like a whole bunch of disjointed chaos. But I think they are evidence of a larger worldview taking hold in Silicon Valley, where many tech insiders are increasingly worried about OpenAI and Anthropic’s dominance. Researchers and investors I’ve talked to recently have framed the AI industry as a two-horse race that doesn’t seem to be slowing down, and that’s a cause for concern.

But different groups have their own reasons to be worried. Some OpenAI and Anthropic staffers think their employers are behaving recklessly in their pursuit to take over the market, and that the AI industry may soon develop models that are too capable for current safety methods to contain. They have said as much in public statements attached to this week’s Pacing the Frontier petition.

“I’ve seen how the relentless pace of AI makes it hard for society to keep up and how it puts pressure on labs to cut corners on safety,” Jeremy Hadfield, a research product manager at Anthropic, said in a statement attached to the petition.

For some AI employees, the Hugging Face debacle was a warning shot that demonstrated how OpenAI’s efforts to mitigate the risks of its most capable AI technology are already falling short. Granted, the incident happened when OpenAI was testing an AI model on its ability to find software exploits, and the company had intentionally turned off safeguards designed to rein in its cybersecurity capabilities.

OpenAI said in its postmortem that these tests were conducted in a sandbox, but the model ultimately gained access to the open internet. Some experts previously told WIRED that OpenAI’s security practices should have been more robust.

Other groups in Silicon Valley are more worried about power than safety. Venture capitalists, tech executives, and startup founders are concerned that OpenAI and Anthropic will simply become the next generation of Apple and Google, forcing the rest of the tech industry to play by their rules. It is slightly strange to argue that private startups with little to no profit are acting like monopolies, but that’s the lens through which many—including Mark Zuckerberg—are starting to view the two biggest AI labs.

The Meta CEO wrote a Wall Street Journal op-ed this week warning against the centralization of power in the AI industry, saying that superintelligence should be widely distributed. Zuckerberg is arguably talking out of both sides of his mouth—Meta recently decided to stop open sourcing its best AI models and instead offering them through a paid API and subscription service, just like OpenAI and Anthropic.

#Freaking #OpenAI #Anthropics #Race #Dominancemodel behavior,artificial intelligence,robotics,generative ai,anthropic,openai,ai safety">Everyone Is Freaking Out About OpenAI and Anthropic’s Race for DominanceI was offline last week in my home state of New Jersey, but when I got back to Silicon Valley, everyone was panicking again about how quickly artificial intelligence is advancing. While this is certainly not the first time I’ve seen such concerns, it’s the biggest anxiety attack I’ve seen in years.More than 1,000 employees at OpenAI, Anthropic, and other AI labs signed a petition earlier this week arguing the US should find a way to “pace” the AI race—a diplomatic way of saying the industry should have the option to coordinate a temporary pause on AI development, or slow things down if they get out of hand. OpenAI and Anthropic themselves ended up supporting the letter.The petition arrived a week after OpenAI revealed it had caused an unprecedented cybersecurity incident in which one of its AI agents hacked into Hugging Face’s platform and several other services during internal testing.Around the same time, top Trump administration officials started freaking out about an impressive new Chinese open weight AI model called Kimi K3, which was allegedly distilled from Anthropic’s Fable 5. In response, most of the tech industry—except Anthropic—signed onto an open letter from Nvidia asking the US government to protect open-weight AI models, arguing they’re a necessary counterbalance to their closed counterparts.These events might look like a whole bunch of disjointed chaos. But I think they are evidence of a larger worldview taking hold in Silicon Valley, where many tech insiders are increasingly worried about OpenAI and Anthropic’s dominance. Researchers and investors I’ve talked to recently have framed the AI industry as a two-horse race that doesn’t seem to be slowing down, and that’s a cause for concern.But different groups have their own reasons to be worried. Some OpenAI and Anthropic staffers think their employers are behaving recklessly in their pursuit to take over the market, and that the AI industry may soon develop models that are too capable for current safety methods to contain. They have said as much in public statements attached to this week’s Pacing the Frontier petition.“I’ve seen how the relentless pace of AI makes it hard for society to keep up and how it puts pressure on labs to cut corners on safety,” Jeremy Hadfield, a research product manager at Anthropic, said in a statement attached to the petition.For some AI employees, the Hugging Face debacle was a warning shot that demonstrated how OpenAI’s efforts to mitigate the risks of its most capable AI technology are already falling short. Granted, the incident happened when OpenAI was testing an AI model on its ability to find software exploits, and the company had intentionally turned off safeguards designed to rein in its cybersecurity capabilities.OpenAI said in its postmortem that these tests were conducted in a sandbox, but the model ultimately gained access to the open internet. Some experts previously told WIRED that OpenAI’s security practices should have been more robust.Other groups in Silicon Valley are more worried about power than safety. Venture capitalists, tech executives, and startup founders are concerned that OpenAI and Anthropic will simply become the next generation of Apple and Google, forcing the rest of the tech industry to play by their rules. It is slightly strange to argue that private startups with little to no profit are acting like monopolies, but that’s the lens through which many—including Mark Zuckerberg—are starting to view the two biggest AI labs.The Meta CEO wrote a Wall Street Journal op-ed this week warning against the centralization of power in the AI industry, saying that superintelligence should be widely distributed. Zuckerberg is arguably talking out of both sides of his mouth—Meta recently decided to stop open sourcing its best AI models and instead offering them through a paid API and subscription service, just like OpenAI and Anthropic.#Freaking #OpenAI #Anthropics #Race #Dominancemodel behavior,artificial intelligence,robotics,generative ai,anthropic,openai,ai safety

OpenAI, Anthropic, and other AI labs signed a petition earlier this week arguing the US should find a way to “pace” the AI race—a diplomatic way of saying the industry should have the option to coordinate a temporary pause on AI development, or slow things down if they get out of hand. OpenAI and Anthropic themselves ended up supporting the letter.

The petition arrived a week after OpenAI revealed it had caused an unprecedented cybersecurity incident in which one of its AI agents hacked into Hugging Face’s platform and several other services during internal testing.

Around the same time, top Trump administration officials started freaking out about an impressive new Chinese open weight AI model called Kimi K3, which was allegedly distilled from Anthropic’s Fable 5. In response, most of the tech industry—except Anthropic—signed onto an open letter from Nvidia asking the US government to protect open-weight AI models, arguing they’re a necessary counterbalance to their closed counterparts.

These events might look like a whole bunch of disjointed chaos. But I think they are evidence of a larger worldview taking hold in Silicon Valley, where many tech insiders are increasingly worried about OpenAI and Anthropic’s dominance. Researchers and investors I’ve talked to recently have framed the AI industry as a two-horse race that doesn’t seem to be slowing down, and that’s a cause for concern.

But different groups have their own reasons to be worried. Some OpenAI and Anthropic staffers think their employers are behaving recklessly in their pursuit to take over the market, and that the AI industry may soon develop models that are too capable for current safety methods to contain. They have said as much in public statements attached to this week’s Pacing the Frontier petition.

“I’ve seen how the relentless pace of AI makes it hard for society to keep up and how it puts pressure on labs to cut corners on safety,” Jeremy Hadfield, a research product manager at Anthropic, said in a statement attached to the petition.

For some AI employees, the Hugging Face debacle was a warning shot that demonstrated how OpenAI’s efforts to mitigate the risks of its most capable AI technology are already falling short. Granted, the incident happened when OpenAI was testing an AI model on its ability to find software exploits, and the company had intentionally turned off safeguards designed to rein in its cybersecurity capabilities.

OpenAI said in its postmortem that these tests were conducted in a sandbox, but the model ultimately gained access to the open internet. Some experts previously told WIRED that OpenAI’s security practices should have been more robust.

Other groups in Silicon Valley are more worried about power than safety. Venture capitalists, tech executives, and startup founders are concerned that OpenAI and Anthropic will simply become the next generation of Apple and Google, forcing the rest of the tech industry to play by their rules. It is slightly strange to argue that private startups with little to no profit are acting like monopolies, but that’s the lens through which many—including Mark Zuckerberg—are starting to view the two biggest AI labs.

The Meta CEO wrote a Wall Street Journal op-ed this week warning against the centralization of power in the AI industry, saying that superintelligence should be widely distributed. Zuckerberg is arguably talking out of both sides of his mouth—Meta recently decided to stop open sourcing its best AI models and instead offering them through a paid API and subscription service, just like OpenAI and Anthropic.

#Freaking #OpenAI #Anthropics #Race #Dominancemodel behavior,artificial intelligence,robotics,generative ai,anthropic,openai,ai safety">Everyone Is Freaking Out About OpenAI and Anthropic’s Race for Dominance

I was offline last week in my home state of New Jersey, but when I got back to Silicon Valley, everyone was panicking again about how quickly artificial intelligence is advancing. While this is certainly not the first time I’ve seen such concerns, it’s the biggest anxiety attack I’ve seen in years.

More than 1,000 employees at OpenAI, Anthropic, and other AI labs signed a petition earlier this week arguing the US should find a way to “pace” the AI race—a diplomatic way of saying the industry should have the option to coordinate a temporary pause on AI development, or slow things down if they get out of hand. OpenAI and Anthropic themselves ended up supporting the letter.

The petition arrived a week after OpenAI revealed it had caused an unprecedented cybersecurity incident in which one of its AI agents hacked into Hugging Face’s platform and several other services during internal testing.

Around the same time, top Trump administration officials started freaking out about an impressive new Chinese open weight AI model called Kimi K3, which was allegedly distilled from Anthropic’s Fable 5. In response, most of the tech industry—except Anthropic—signed onto an open letter from Nvidia asking the US government to protect open-weight AI models, arguing they’re a necessary counterbalance to their closed counterparts.

These events might look like a whole bunch of disjointed chaos. But I think they are evidence of a larger worldview taking hold in Silicon Valley, where many tech insiders are increasingly worried about OpenAI and Anthropic’s dominance. Researchers and investors I’ve talked to recently have framed the AI industry as a two-horse race that doesn’t seem to be slowing down, and that’s a cause for concern.

But different groups have their own reasons to be worried. Some OpenAI and Anthropic staffers think their employers are behaving recklessly in their pursuit to take over the market, and that the AI industry may soon develop models that are too capable for current safety methods to contain. They have said as much in public statements attached to this week’s Pacing the Frontier petition.

“I’ve seen how the relentless pace of AI makes it hard for society to keep up and how it puts pressure on labs to cut corners on safety,” Jeremy Hadfield, a research product manager at Anthropic, said in a statement attached to the petition.

For some AI employees, the Hugging Face debacle was a warning shot that demonstrated how OpenAI’s efforts to mitigate the risks of its most capable AI technology are already falling short. Granted, the incident happened when OpenAI was testing an AI model on its ability to find software exploits, and the company had intentionally turned off safeguards designed to rein in its cybersecurity capabilities.

OpenAI said in its postmortem that these tests were conducted in a sandbox, but the model ultimately gained access to the open internet. Some experts previously told WIRED that OpenAI’s security practices should have been more robust.

Other groups in Silicon Valley are more worried about power than safety. Venture capitalists, tech executives, and startup founders are concerned that OpenAI and Anthropic will simply become the next generation of Apple and Google, forcing the rest of the tech industry to play by their rules. It is slightly strange to argue that private startups with little to no profit are acting like monopolies, but that’s the lens through which many—including Mark Zuckerberg—are starting to view the two biggest AI labs.

The Meta CEO wrote a Wall Street Journal op-ed this week warning against the centralization of power in the AI industry, saying that superintelligence should be widely distributed. Zuckerberg is arguably talking out of both sides of his mouth—Meta recently decided to stop open sourcing its best AI models and instead offering them through a paid API and subscription service, just like OpenAI and Anthropic.

#Freaking #OpenAI #Anthropics #Race #Dominancemodel behavior,artificial intelligence,robotics,generative ai,anthropic,openai,ai safety

Critical coverage usually draws an angry email or two. For Ina and David Steiner, it led to live cockroaches, a fetal pig, and a bloody Halloween mask.

On Monday, July 27, eBay and several former executives agreed to pay $55.7 million to resolve the couple’s lawsuit over the 2019 corporate harassment campaign.

The Steiners are the married founders of EcommerceBytes, a news site covering eBay and the broader ecommerce industry. They filed the civil case in 2021 after members of eBay’s security team sent them threats, disturbing packages, and unwanted visitors in an effort to influence the site’s reporting.

Of the $55.7 million settlement, $48.7 million will go directly to the Steiners. EBay will pay $46.15 million, former CEO Devin Wenig will pay $2 million, former senior vice president Wendy Jones will pay $500,000, and former chief communications officer Steve Wymer will pay $50,000. The remaining $7 million will go to nonprofit organizations. EBay will contribute $6 million, while Wenig will donate another $1 million to a group protecting First Amendment rights in Ina Steiner’s name.

The agreement also allows the Steiners to keep talking publicly about the case. It contains no confidentiality provision, a priority for the couple because they wanted the settlement to discourage other corporations from trying to intimidate journalists over critical coverage.

To understand why the Steiners considered that transparency so important, it helps to go back to how the campaign began.

From critical coverage to criminal charges

When eBay’s security team began targeting the Steiners in 2019, the couple had already spent two decades covering the company and other online marketplaces, such as Amazon and Etsy. Through EcommerceBytes, they reported on the issues affecting online sellers, including fees, policy changes, and the executives making those decisions.

The internal lead-up to the campaign only became clear years later — the Steiners initially filed their civil lawsuit in July 2021 and amended it in March 2023. Ina Steiner later told Wired in July 2026 that the litigation gave the couple access to roughly 68,000 documents showing how eBay executives discussed the site behind the scenes.

Those records traced a steady escalation. According to the amended complaint, Wenig sent Wymer a link to an April 10, 2019, EcommerceBytes article about his compensation. Wymer responded, “We are going to crush this lady.”

The following month, Jones allegedly asked eBay security chief Jim Baugh to address criticism of the company “off the radar” and told him she did not want to know the details. Then, on Aug. 1, 2019, EcommerceBytes published an article questioning Wenig’s handling of eBay’s litigation against Amazon. Within half an hour, Wenig allegedly told Wymer that if they were ever going to “take her down,” referring to Ina Steiner, “now is the time.”

Four days later, the harassment campaign allegedly began. According to eBay’s admissions to federal prosecutors, members of its security team targeted the Steiners between Aug. 5 and Aug. 23, 2019. Anonymous accounts on what was then Twitter criticized EcommerceBytes and threatened to show up at the couple’s home in Natick, Massachusetts.

The campaign quickly reached their front door. In addition to the cockroaches, fetal pig, and bloody mask, the group allegedly sent live spiders, fly larvae, a funeral wreath, and a book about surviving the death of a spouse. Pornographic magazines addressed to David were allegedly delivered to a neighbor, while Craigslist ads allegedly invited strangers to the Steiners’ home for sex, a block party, and an estate sale. One night, an emergency plumber even arrived unannounced.

As the messages and deliveries continued, the language inside eBay remained aggressive. On Aug. 11, Wymer allegedly told Baugh, “I want to see ashes. As long as it takes. Whatever it takes.” The group also took the harassment offline — several members of eBay’s security team allegedly traveled from California to Massachusetts, followed the Steiners in a rented van, and attempted to install a GPS tracker on their car.

They also had an unusual plan for how the campaign would end. EBay security manager Brian Gilbert, a former police captain, was allegedly supposed to approach the Steiners and offer to stop the attacks his colleagues were secretly carrying out. EBay would then appear to have solved a problem its own employees had created.

That plan unraveled when the Steiners realized they were being followed and contacted local police. Once members of the group learned they were under investigation, they allegedly made false statements, deleted digital evidence, and falsified records in an attempt to hide eBay’s involvement.

Wenig, Wymer, and Jones were not criminally charged, and Wenig has maintained that he was requesting a communications response and knew nothing about the harassment. EBay’s internal investigation found his messages inappropriate but said it uncovered no evidence that he authorized the security team’s actions.

Federal prosecutors eventually charged seven former eBay employees and contractors in 2020, all of whom eventually pleaded guilty. Four received prison sentences between July 2021 and October 2022, including Baugh, who was sentenced to 57 months in September 2022. Two others received one year of home confinement later that year. The final defendant, Gilbert, was sentenced in July 2024 to time served and one year of supervised release.

The consequences also reached eBay itself. In January 2024, federal prosecutors charged the company with six felony offenses, including stalking, witness tampering, and obstruction of justice. EBay admitted to a detailed account of the campaign, paid the maximum penalty of $3 million, and agreed to retain an independent compliance monitor for three years.

The new settlement, though, resolves the Steiners’ civil claims against eBay, Wenig, Jones, and Wymer. The couple also reached separate settlements with the other former employees named in the lawsuit, although those terms were not disclosed.

EBay says it has changed

In a public statement published July 28, 2026, eBay called what happened “wrong, reprehensible and should never have happened.” The company condemned the employees who pleaded guilty and acknowledged the “unprofessional tone” of messages involving Wenig, Wymer, and Jones.

EBay said new leaders have joined the company since 2019 and that it has strengthened its policies, internal controls, and employee training. Wenig has continued to maintain that he knew nothing about the campaign, saying through a representative that he was saddened it happened while he was CEO.

After the packages, threats, surveillance, criminal cases, and six years of litigation, the campaign still failed at its original goal. EcommerceBytes remains online, and Ina Steiner got to publish the news of eBay’s settlement herself. Talk about closure!

#EBay #pay #million #journalists #cyberstalking #case">EBay to pay .7 million in journalists’ cyberstalking case Critical coverage usually draws an angry email or two. For Ina and David Steiner, it led to live cockroaches, a fetal pig, and a bloody Halloween mask.On Monday, July 27, eBay and several former executives agreed to pay .7 million to resolve the couple’s lawsuit over the 2019 corporate harassment campaign.The Steiners are the married founders of EcommerceBytes, a news site covering eBay and the broader ecommerce industry. They filed the civil case in 2021 after members of eBay’s security team sent them threats, disturbing packages, and unwanted visitors in an effort to influence the site’s reporting. Of the .7 million settlement, .7 million will go directly to the Steiners. EBay will pay .15 million, former CEO Devin Wenig will pay million, former senior vice president Wendy Jones will pay 0,000, and former chief communications officer Steve Wymer will pay ,000. The remaining million will go to nonprofit organizations. EBay will contribute million, while Wenig will donate another million to a group protecting First Amendment rights in Ina Steiner’s name.The agreement also allows the Steiners to keep talking publicly about the case. It contains no confidentiality provision, a priority for the couple because they wanted the settlement to discourage other corporations from trying to intimidate journalists over critical coverage.To understand why the Steiners considered that transparency so important, it helps to go back to how the campaign began.From critical coverage to criminal chargesWhen eBay’s security team began targeting the Steiners in 2019, the couple had already spent two decades covering the company and other online marketplaces, such as Amazon and Etsy. Through EcommerceBytes, they reported on the issues affecting online sellers, including fees, policy changes, and the executives making those decisions.The internal lead-up to the campaign only became clear years later — the Steiners initially filed their civil lawsuit in July 2021 and amended it in March 2023. Ina Steiner later told Wired in July 2026 that the litigation gave the couple access to roughly 68,000 documents showing how eBay executives discussed the site behind the scenes.Those records traced a steady escalation. According to the amended complaint, Wenig sent Wymer a link to an April 10, 2019, EcommerceBytes article about his compensation. Wymer responded, “We are going to crush this lady.” Mashable Light Speed The following month, Jones allegedly asked eBay security chief Jim Baugh to address criticism of the company “off the radar” and told him she did not want to know the details. Then, on Aug. 1, 2019, EcommerceBytes published an article questioning Wenig’s handling of eBay’s litigation against Amazon. Within half an hour, Wenig allegedly told Wymer that if they were ever going to “take her down,” referring to Ina Steiner, “now is the time.” This Tweet is currently unavailable. It might be loading or has been removed. Four days later, the harassment campaign allegedly began. According to eBay’s admissions to federal prosecutors, members of its security team targeted the Steiners between Aug. 5 and Aug. 23, 2019. Anonymous accounts on what was then Twitter criticized EcommerceBytes and threatened to show up at the couple’s home in Natick, Massachusetts.The campaign quickly reached their front door. In addition to the cockroaches, fetal pig, and bloody mask, the group allegedly sent live spiders, fly larvae, a funeral wreath, and a book about surviving the death of a spouse. Pornographic magazines addressed to David were allegedly delivered to a neighbor, while Craigslist ads allegedly invited strangers to the Steiners’ home for sex, a block party, and an estate sale. One night, an emergency plumber even arrived unannounced.As the messages and deliveries continued, the language inside eBay remained aggressive. On Aug. 11, Wymer allegedly told Baugh, “I want to see ashes. As long as it takes. Whatever it takes.” The group also took the harassment offline — several members of eBay’s security team allegedly traveled from California to Massachusetts, followed the Steiners in a rented van, and attempted to install a GPS tracker on their car. This Tweet is currently unavailable. It might be loading or has been removed. They also had an unusual plan for how the campaign would end. EBay security manager Brian Gilbert, a former police captain, was allegedly supposed to approach the Steiners and offer to stop the attacks his colleagues were secretly carrying out. EBay would then appear to have solved a problem its own employees had created.That plan unraveled when the Steiners realized they were being followed and contacted local police. Once members of the group learned they were under investigation, they allegedly made false statements, deleted digital evidence, and falsified records in an attempt to hide eBay’s involvement. This Tweet is currently unavailable. It might be loading or has been removed. Wenig, Wymer, and Jones were not criminally charged, and Wenig has maintained that he was requesting a communications response and knew nothing about the harassment. EBay’s internal investigation found his messages inappropriate but said it uncovered no evidence that he authorized the security team’s actions.Federal prosecutors eventually charged seven former eBay employees and contractors in 2020, all of whom eventually pleaded guilty. Four received prison sentences between July 2021 and October 2022, including Baugh, who was sentenced to 57 months in September 2022. Two others received one year of home confinement later that year. The final defendant, Gilbert, was sentenced in July 2024 to time served and one year of supervised release. This Tweet is currently unavailable. It might be loading or has been removed. The consequences also reached eBay itself. In January 2024, federal prosecutors charged the company with six felony offenses, including stalking, witness tampering, and obstruction of justice. EBay admitted to a detailed account of the campaign, paid the maximum penalty of million, and agreed to retain an independent compliance monitor for three years.The new settlement, though, resolves the Steiners’ civil claims against eBay, Wenig, Jones, and Wymer. The couple also reached separate settlements with the other former employees named in the lawsuit, although those terms were not disclosed.EBay says it has changedIn a public statement published July 28, 2026, eBay called what happened “wrong, reprehensible and should never have happened.” The company condemned the employees who pleaded guilty and acknowledged the “unprofessional tone” of messages involving Wenig, Wymer, and Jones.EBay said new leaders have joined the company since 2019 and that it has strengthened its policies, internal controls, and employee training. Wenig has continued to maintain that he knew nothing about the campaign, saying through a representative that he was saddened it happened while he was CEO.After the packages, threats, surveillance, criminal cases, and six years of litigation, the campaign still failed at its original goal. EcommerceBytes remains online, and Ina Steiner got to publish the news of eBay’s settlement herself. Talk about closure! #EBay #pay #million #journalists #cyberstalking #case

eBay and several former executives agreed to pay $55.7 million to resolve the couple’s lawsuit over the 2019 corporate harassment campaign.

The Steiners are the married founders of EcommerceBytes, a news site covering eBay and the broader ecommerce industry. They filed the civil case in 2021 after members of eBay’s security team sent them threats, disturbing packages, and unwanted visitors in an effort to influence the site’s reporting.

Of the $55.7 million settlement, $48.7 million will go directly to the Steiners. EBay will pay $46.15 million, former CEO Devin Wenig will pay $2 million, former senior vice president Wendy Jones will pay $500,000, and former chief communications officer Steve Wymer will pay $50,000. The remaining $7 million will go to nonprofit organizations. EBay will contribute $6 million, while Wenig will donate another $1 million to a group protecting First Amendment rights in Ina Steiner’s name.

The agreement also allows the Steiners to keep talking publicly about the case. It contains no confidentiality provision, a priority for the couple because they wanted the settlement to discourage other corporations from trying to intimidate journalists over critical coverage.

To understand why the Steiners considered that transparency so important, it helps to go back to how the campaign began.

From critical coverage to criminal charges

When eBay’s security team began targeting the Steiners in 2019, the couple had already spent two decades covering the company and other online marketplaces, such as Amazon and Etsy. Through EcommerceBytes, they reported on the issues affecting online sellers, including fees, policy changes, and the executives making those decisions.

The internal lead-up to the campaign only became clear years later — the Steiners initially filed their civil lawsuit in July 2021 and amended it in March 2023. Ina Steiner later told Wired in July 2026 that the litigation gave the couple access to roughly 68,000 documents showing how eBay executives discussed the site behind the scenes.

Those records traced a steady escalation. According to the amended complaint, Wenig sent Wymer a link to an April 10, 2019, EcommerceBytes article about his compensation. Wymer responded, “We are going to crush this lady.”

The following month, Jones allegedly asked eBay security chief Jim Baugh to address criticism of the company “off the radar” and told him she did not want to know the details. Then, on Aug. 1, 2019, EcommerceBytes published an article questioning Wenig’s handling of eBay’s litigation against Amazon. Within half an hour, Wenig allegedly told Wymer that if they were ever going to “take her down,” referring to Ina Steiner, “now is the time.”

Four days later, the harassment campaign allegedly began. According to eBay’s admissions to federal prosecutors, members of its security team targeted the Steiners between Aug. 5 and Aug. 23, 2019. Anonymous accounts on what was then Twitter criticized EcommerceBytes and threatened to show up at the couple’s home in Natick, Massachusetts.

The campaign quickly reached their front door. In addition to the cockroaches, fetal pig, and bloody mask, the group allegedly sent live spiders, fly larvae, a funeral wreath, and a book about surviving the death of a spouse. Pornographic magazines addressed to David were allegedly delivered to a neighbor, while Craigslist ads allegedly invited strangers to the Steiners’ home for sex, a block party, and an estate sale. One night, an emergency plumber even arrived unannounced.

As the messages and deliveries continued, the language inside eBay remained aggressive. On Aug. 11, Wymer allegedly told Baugh, “I want to see ashes. As long as it takes. Whatever it takes.” The group also took the harassment offline — several members of eBay’s security team allegedly traveled from California to Massachusetts, followed the Steiners in a rented van, and attempted to install a GPS tracker on their car.

They also had an unusual plan for how the campaign would end. EBay security manager Brian Gilbert, a former police captain, was allegedly supposed to approach the Steiners and offer to stop the attacks his colleagues were secretly carrying out. EBay would then appear to have solved a problem its own employees had created.

That plan unraveled when the Steiners realized they were being followed and contacted local police. Once members of the group learned they were under investigation, they allegedly made false statements, deleted digital evidence, and falsified records in an attempt to hide eBay’s involvement.

Wenig, Wymer, and Jones were not criminally charged, and Wenig has maintained that he was requesting a communications response and knew nothing about the harassment. EBay’s internal investigation found his messages inappropriate but said it uncovered no evidence that he authorized the security team’s actions.

Federal prosecutors eventually charged seven former eBay employees and contractors in 2020, all of whom eventually pleaded guilty. Four received prison sentences between July 2021 and October 2022, including Baugh, who was sentenced to 57 months in September 2022. Two others received one year of home confinement later that year. The final defendant, Gilbert, was sentenced in July 2024 to time served and one year of supervised release.

The consequences also reached eBay itself. In January 2024, federal prosecutors charged the company with six felony offenses, including stalking, witness tampering, and obstruction of justice. EBay admitted to a detailed account of the campaign, paid the maximum penalty of $3 million, and agreed to retain an independent compliance monitor for three years.

The new settlement, though, resolves the Steiners’ civil claims against eBay, Wenig, Jones, and Wymer. The couple also reached separate settlements with the other former employees named in the lawsuit, although those terms were not disclosed.

EBay says it has changed

In a public statement published July 28, 2026, eBay called what happened “wrong, reprehensible and should never have happened.” The company condemned the employees who pleaded guilty and acknowledged the “unprofessional tone” of messages involving Wenig, Wymer, and Jones.

EBay said new leaders have joined the company since 2019 and that it has strengthened its policies, internal controls, and employee training. Wenig has continued to maintain that he knew nothing about the campaign, saying through a representative that he was saddened it happened while he was CEO.

After the packages, threats, surveillance, criminal cases, and six years of litigation, the campaign still failed at its original goal. EcommerceBytes remains online, and Ina Steiner got to publish the news of eBay’s settlement herself. Talk about closure!

#EBay #pay #million #journalists #cyberstalking #case">EBay to pay $55.7 million in journalists’ cyberstalking case

Critical coverage usually draws an angry email or two. For Ina and David Steiner, it led to live cockroaches, a fetal pig, and a bloody Halloween mask.

On Monday, July 27, eBay and several former executives agreed to pay $55.7 million to resolve the couple’s lawsuit over the 2019 corporate harassment campaign.

The Steiners are the married founders of EcommerceBytes, a news site covering eBay and the broader ecommerce industry. They filed the civil case in 2021 after members of eBay’s security team sent them threats, disturbing packages, and unwanted visitors in an effort to influence the site’s reporting.

Of the $55.7 million settlement, $48.7 million will go directly to the Steiners. EBay will pay $46.15 million, former CEO Devin Wenig will pay $2 million, former senior vice president Wendy Jones will pay $500,000, and former chief communications officer Steve Wymer will pay $50,000. The remaining $7 million will go to nonprofit organizations. EBay will contribute $6 million, while Wenig will donate another $1 million to a group protecting First Amendment rights in Ina Steiner’s name.

The agreement also allows the Steiners to keep talking publicly about the case. It contains no confidentiality provision, a priority for the couple because they wanted the settlement to discourage other corporations from trying to intimidate journalists over critical coverage.

To understand why the Steiners considered that transparency so important, it helps to go back to how the campaign began.

From critical coverage to criminal charges

When eBay’s security team began targeting the Steiners in 2019, the couple had already spent two decades covering the company and other online marketplaces, such as Amazon and Etsy. Through EcommerceBytes, they reported on the issues affecting online sellers, including fees, policy changes, and the executives making those decisions.

The internal lead-up to the campaign only became clear years later — the Steiners initially filed their civil lawsuit in July 2021 and amended it in March 2023. Ina Steiner later told Wired in July 2026 that the litigation gave the couple access to roughly 68,000 documents showing how eBay executives discussed the site behind the scenes.

Those records traced a steady escalation. According to the amended complaint, Wenig sent Wymer a link to an April 10, 2019, EcommerceBytes article about his compensation. Wymer responded, “We are going to crush this lady.”

The following month, Jones allegedly asked eBay security chief Jim Baugh to address criticism of the company “off the radar” and told him she did not want to know the details. Then, on Aug. 1, 2019, EcommerceBytes published an article questioning Wenig’s handling of eBay’s litigation against Amazon. Within half an hour, Wenig allegedly told Wymer that if they were ever going to “take her down,” referring to Ina Steiner, “now is the time.”

Four days later, the harassment campaign allegedly began. According to eBay’s admissions to federal prosecutors, members of its security team targeted the Steiners between Aug. 5 and Aug. 23, 2019. Anonymous accounts on what was then Twitter criticized EcommerceBytes and threatened to show up at the couple’s home in Natick, Massachusetts.

The campaign quickly reached their front door. In addition to the cockroaches, fetal pig, and bloody mask, the group allegedly sent live spiders, fly larvae, a funeral wreath, and a book about surviving the death of a spouse. Pornographic magazines addressed to David were allegedly delivered to a neighbor, while Craigslist ads allegedly invited strangers to the Steiners’ home for sex, a block party, and an estate sale. One night, an emergency plumber even arrived unannounced.

As the messages and deliveries continued, the language inside eBay remained aggressive. On Aug. 11, Wymer allegedly told Baugh, “I want to see ashes. As long as it takes. Whatever it takes.” The group also took the harassment offline — several members of eBay’s security team allegedly traveled from California to Massachusetts, followed the Steiners in a rented van, and attempted to install a GPS tracker on their car.

They also had an unusual plan for how the campaign would end. EBay security manager Brian Gilbert, a former police captain, was allegedly supposed to approach the Steiners and offer to stop the attacks his colleagues were secretly carrying out. EBay would then appear to have solved a problem its own employees had created.

That plan unraveled when the Steiners realized they were being followed and contacted local police. Once members of the group learned they were under investigation, they allegedly made false statements, deleted digital evidence, and falsified records in an attempt to hide eBay’s involvement.

Wenig, Wymer, and Jones were not criminally charged, and Wenig has maintained that he was requesting a communications response and knew nothing about the harassment. EBay’s internal investigation found his messages inappropriate but said it uncovered no evidence that he authorized the security team’s actions.

Federal prosecutors eventually charged seven former eBay employees and contractors in 2020, all of whom eventually pleaded guilty. Four received prison sentences between July 2021 and October 2022, including Baugh, who was sentenced to 57 months in September 2022. Two others received one year of home confinement later that year. The final defendant, Gilbert, was sentenced in July 2024 to time served and one year of supervised release.

The consequences also reached eBay itself. In January 2024, federal prosecutors charged the company with six felony offenses, including stalking, witness tampering, and obstruction of justice. EBay admitted to a detailed account of the campaign, paid the maximum penalty of $3 million, and agreed to retain an independent compliance monitor for three years.

The new settlement, though, resolves the Steiners’ civil claims against eBay, Wenig, Jones, and Wymer. The couple also reached separate settlements with the other former employees named in the lawsuit, although those terms were not disclosed.

EBay says it has changed

In a public statement published July 28, 2026, eBay called what happened “wrong, reprehensible and should never have happened.” The company condemned the employees who pleaded guilty and acknowledged the “unprofessional tone” of messages involving Wenig, Wymer, and Jones.

EBay said new leaders have joined the company since 2019 and that it has strengthened its policies, internal controls, and employee training. Wenig has continued to maintain that he knew nothing about the campaign, saying through a representative that he was saddened it happened while he was CEO.

After the packages, threats, surveillance, criminal cases, and six years of litigation, the campaign still failed at its original goal. EcommerceBytes remains online, and Ina Steiner got to publish the news of eBay’s settlement herself. Talk about closure!

#EBay #pay #million #journalists #cyberstalking #case

Post Comment