home.social

#safeguards — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #safeguards, aggregated by home.social.

fetched live
  1. Access to #ClaudeFable 5 and #ClaudeMythos 5 has been #restored after the #USgovernment lifted #exportcontrols. The controls were imposed due to a report of a method bypassing #Fable 5’s #safeguards, which has now been addressed with an improved safety classifier. #Anthropic is collaborating with the US government and industry partners to develop a shared framework for assessing and mitigating AI model “jailbreaks.” anthropic.com/news/redeploying #tech #media #news

  2. Access to #ClaudeFable 5 and #ClaudeMythos 5 has been #restored after the #USgovernment lifted #exportcontrols. The controls were imposed due to a report of a method bypassing #Fable 5’s #safeguards, which has now been addressed with an improved safety classifier. #Anthropic is collaborating with the US government and industry partners to develop a shared framework for assessing and mitigating AI model “jailbreaks.” anthropic.com/news/redeploying #tech #media #news

  3. What's important with handling AI is to ensure the right framework is in place for proper safeguards.

    Humans need to be in place to ensure that whatever AI is producing and is responsible for is correct and ethical. AI can not be held accountable.
    #ai #safeguards #framework #safety #ethics

  4. What's important with handling AI is to ensure the right framework is in place for proper safeguards.

    Humans need to be in place to ensure that whatever AI is producing and is responsible for is correct and ethical. AI can not be held accountable.
    #ai #safeguards #framework #safety #ethics

  5. Another test lawsuit launched accusing ChatGPT of failing to detect threats of possible violence or crime.

    OpenAI sued by family of shooting victim in the April 2025 mass shooting at Florida State University. The lawsuit alleges that OpenAI’s ChatGPT enabled the attack. nbcnews.com/news/us-news/opena #AI #ChatGPT #OpenAI #Violence #Crime #LLMs #LocusofResponsibility #Lawsuit #SafeGuards #MentalHealth #Investigation #MassShooting

  6. Another test lawsuit launched accusing ChatGPT of failing to detect threats of possible violence or crime.

    OpenAI sued by family of shooting victim in the April 2025 mass shooting at Florida State University. The lawsuit alleges that OpenAI’s ChatGPT enabled the attack. nbcnews.com/news/us-news/opena

  7. #Japan lifts curbs on #weapons #exports, but strict #safeguards are essential
    "If te #defenseindustry is treated as an engine of economic growth, it becomes too easy to fall into a profit-first mindset.. At its core, arms exports mean 🇯🇵becoming involved, even indirectly, in other countries' #conflicts.
    To determine whether tt truly contributes to 🇯🇵's security, #government must fulfill its #responsibility to explain decisions to te public on a concrete, case-by-case basis"
    english.kyodonews.net/articles

  8. #Japan lifts curbs on #weapons #exports, but strict #safeguards are essential
    "If te #defenseindustry is treated as an engine of economic growth, it becomes too easy to fall into a profit-first mindset.. At its core, arms exports mean 🇯🇵becoming involved, even indirectly, in other countries' #conflicts.
    To determine whether tt truly contributes to 🇯🇵's security, #government must fulfill its #responsibility to explain decisions to te public on a concrete, case-by-case basis"
    english.kyodonews.net/articles

  9. #OpenAI released #policyproposals for managing the #economicimpact of #AI, like shifting the tax burden from #labour to #capital, implementing a #robottax, and creating a #PublicWealthFund. The proposals also suggest labour-focused measures like a subsidised four-day work week. OpenAI emphasises the need for #safeguards against #AIrisks. techcrunch.com/2026/04/06/open #tech #media #news

  10. #OpenAI released #policyproposals for managing the #economicimpact of #AI, like shifting the tax burden from #labour to #capital, implementing a #robottax, and creating a #PublicWealthFund. The proposals also suggest labour-focused measures like a subsidised four-day work week. OpenAI emphasises the need for #safeguards against #AIrisks. techcrunch.com/2026/04/06/open #tech #media #news

  11. #Anthropic sfida il #governoUSA in #tribunale per bloccare la sua inclusione nella #listanera per la #sicurezzanazionale. La disputa riguarda le #safeguards etiche sui suoi modelli di #IA e il rischio percepito per la catena di approvvigionamento. Per #investitori e settore #tech, da seguire: impatto su contratti governativi, precedenti legali e il futuro di AI e #quantumcomputing
    @attualita
    @economia

  12. #Anthropic sfida il #governoUSA in #tribunale per bloccare la sua inclusione nella #listanera per la #sicurezzanazionale. La disputa riguarda le #safeguards etiche sui suoi modelli di #IA e il rischio percepito per la catena di approvvigionamento. Per #investitori e settore #tech, da seguire: impatto su contratti governativi, precedenti legali e il futuro di AI e #quantumcomputing
    @attualita
    @economia

  13. A #study testing 13 #LLMs found that all can be used to commit #academicfraud or facilitate #junkscience. While some models, like Claude, were more resistant to fraudulent requests, others, like Grok and early GPT versions, performed poorly. The study highlights the need for developers to implement stronger #safeguards against #misuse of LLMs. nature.com/articles/d41586-026 #tech #media #news

  14. A #study testing 13 #LLMs found that all can be used to commit #academicfraud or facilitate #junkscience. While some models, like Claude, were more resistant to fraudulent requests, others, like Grok and early GPT versions, performed poorly. The study highlights the need for developers to implement stronger #safeguards against #misuse of LLMs. nature.com/articles/d41586-026 #tech #media #news

  15. While AI leaders might talk about safeguards, the only ones they have implemented so far are those to safeguard their personal income. Everything else, from regulation to expert opinions, gets decried as negative talk.

    manilatimes.net/2026/03/08/bus

  16. While AI leaders might talk about safeguards, the only ones they have implemented so far are those to safeguard their personal income. Everything else, from regulation to expert opinions, gets decried as negative talk.
    #AI #safeguards
    manilatimes.net/2026/03/08/bus

  17. disabilitynewsservice.com/disa. "The bill’s opponents - including Baroness Grey-Thompson - say their #amendments are crucial to add #safeguards to a private members’ bill that is vague, unsafe & poorly-drafted."

  18. disabilitynewsservice.com/disa. "The bill’s opponents - including Baroness Grey-Thompson - say their #amendments are crucial to add #safeguards to a private members’ bill that is vague, unsafe & poorly-drafted."

  19. Anthropic refuses to bend to Pentagon on AI safeguards as dispute nears deadline

    misryoum.com/us/us-news/anthro

    A public showdown between the Trump administration and Anthropic is hitting an impasse as military officials demand the artificial intelligence company bend its ethical policies by Friday or risk damaging its business.Anthropic CEO Dario Amodei drew a sharp red line...

    #Anthropic #refuses #bend #Pentagon #safeguards #dispute #nears #deadline #US_News_Hub #misryoum_com

  20. #Anthropic is in a dispute with the #Pentagon over its usage restrictions for military purposes. The Pentagon wants Anthropic to remove #safeguards preventing its technology from being used for #autonomousweapons targeting and #domesticsurveillance. Anthropic refuses, citing concerns about responsible use, and the Pentagon has threatened to label it a supply-chain risk. reuters.com/world/anthropic-di #tech #media #news

  21. #Anthropic is in a dispute with the #Pentagon over its usage restrictions for military purposes. The Pentagon wants Anthropic to remove #safeguards preventing its technology from being used for #autonomousweapons targeting and #domesticsurveillance. Anthropic refuses, citing concerns about responsible use, and the Pentagon has threatened to label it a supply-chain risk. reuters.com/world/anthropic-di #tech #media #news

  22. HT @rmblaber1956

    #EnvironmentalGroups sue #Trump’s #EPA over repeal of landmark #climate finding

    Lawsuit from #health and #EnvironmentalJustice groups challenges the EPA’s rollback of the ‘endangerment finding’

    Dharna Noor
    Wed 18 Feb 2026 07.10 EST

    "More than a dozen health and environmental justice non-profits have sued the Environmental Protection Agency over its revocation of the legal determination that underpins US federal climate regulations.

    "Filed in Washington DC circuit court, the lawsuit challenges the EPA’s rollback of the '#EndangermentFinding', which states that the buildup of heat-trapping #pollution in the atmosphere endangers #PublicHealth and welfare and has allowed the EPA to limit those emissions from #vehicles, #PowerPlants and other #industrial sources since 2009. The rollback was widely seen as a major setback to US efforts to combat the #ClimateCrisis.

    "The suit was brought by the American Public Health Association [#APHA], #AmericanLungAssociation, the #CenterForBiologicalDiversity, the #EnvironmentalDefenseFund, the #NaturalResourcesDefenseCouncil [#NRDC], the #SierraClub and 11 other public health and environmental organizations. The lawsuit was filed by green legal organizations #CleanAirTaskForce and #Earthjustice and it names the EPA and the agency’s administrator, #LeeZeldin, as defendants.

    " 'EPA’s repeal of the endangerment finding and #safeguards to limit vehicle emissions marks a complete dereliction of the agency’s mission to protect people’s health and its legal obligation under the #CleanAirAct,' said #GretchenGoldman, president and CEO at the #UnionOfConcernedScientists, another one of the groups behind the lawsuit. 'This shameful and dangerous action by the Trump administration and EPA Administrator Zeldin is rooted in falsehoods not facts and is at complete odds with the public interest and the best available science.' "

    Read more:
    theguardian.com/us-news/2026/f

    Archived version:
    archive.ph/kp4pY

    #USPol #ClimateChange #EPAFail #ClimateCrisis #GreenhouseGas #FossilFools #FossilFuelIndustry #BigOilAndGas #CorporateColonialism #Oiligarchy #Oligarchy #CorporatePolluters

  23. HT @rmblaber1956

    #EnvironmentalGroups sue #Trump’s #EPA over repeal of landmark #climate finding

    Lawsuit from #health and #EnvironmentalJustice groups challenges the EPA’s rollback of the ‘endangerment finding’

    Dharna Noor
    Wed 18 Feb 2026 07.10 EST

    "More than a dozen health and environmental justice non-profits have sued the Environmental Protection Agency over its revocation of the legal determination that underpins US federal climate regulations.

    "Filed in Washington DC circuit court, the lawsuit challenges the EPA’s rollback of the '#EndangermentFinding', which states that the buildup of heat-trapping #pollution in the atmosphere endangers #PublicHealth and welfare and has allowed the EPA to limit those emissions from #vehicles, #PowerPlants and other #industrial sources since 2009. The rollback was widely seen as a major setback to US efforts to combat the #ClimateCrisis.

    "The suit was brought by the American Public Health Association [#APHA], #AmericanLungAssociation, the #CenterForBiologicalDiversity, the #EnvironmentalDefenseFund, the #NaturalResourcesDefenseCouncil [#NRDC], the #SierraClub and 11 other public health and environmental organizations. The lawsuit was filed by green legal organizations #CleanAirTaskForce and #Earthjustice and it names the EPA and the agency’s administrator, #LeeZeldin, as defendants.

    " 'EPA’s repeal of the endangerment finding and #safeguards to limit vehicle emissions marks a complete dereliction of the agency’s mission to protect people’s health and its legal obligation under the #CleanAirAct,' said #GretchenGoldman, president and CEO at the #UnionOfConcernedScientists, another one of the groups behind the lawsuit. 'This shameful and dangerous action by the Trump administration and EPA Administrator Zeldin is rooted in falsehoods not facts and is at complete odds with the public interest and the best available science.' "

    Read more:
    theguardian.com/us-news/2026/f

    Archived version:
    archive.ph/kp4pY

    #USPol #ClimateChange #EPAFail #ClimateCrisis #GreenhouseGas #FossilFools #FossilFuelIndustry #BigOilAndGas #CorporateColonialism #Oiligarchy #Oligarchy #CorporatePolluters

  24. #YouTube CEO #NealMohan’s 2026 priorities include supporting #creators in building sustainable #businesses, leveraging #AI for innovation while combating low-quality content (“#AIslop”), and enhancing #safeguards for #children. Mohan emphasises the importance of transparency and protections regarding AI-generated content, particularly deepfakes. hollywoodreporter.com/business #tech #media #news

  25. #YouTube CEO #NealMohan’s 2026 priorities include supporting #creators in building sustainable #businesses, leveraging #AI for innovation while combating low-quality content (“#AIslop”), and enhancing #safeguards for #children. Mohan emphasises the importance of transparency and protections regarding AI-generated content, particularly deepfakes. hollywoodreporter.com/business #tech #media #news

  26. #Anthropic has implemented technical #safeguards to prevent #thirdparty applications from accessing its #Claude #AImodels through unauthorised means, disrupting workflows for users of tools like OpenCode. This move aims to funnel high-volume automation towards sanctioned channels like the #CommercialAPI or #ClaudeCode. venturebeat.com/technology/ant #tech #media #news

  27. #Anthropic has implemented technical #safeguards to prevent #thirdparty applications from accessing its #Claude #AImodels through unauthorised means, disrupting workflows for users of tools like OpenCode. This move aims to funnel high-volume automation towards sanctioned channels like the #CommercialAPI or #ClaudeCode. venturebeat.com/technology/ant #tech #media #news

  28. danielpinchbeck.substack.com/p

    Toxic Avengers
    From #Paraquat to #PFAS, a systematic dismantling of #environmental #safeguards reveals a disturbing end-times fanaticism
    #DanielPinchbeck

    ..."Since the 1960s, when activists and scientists like Rachel Carson started to realize we were annihilating the biological basis of our future survival with pesticides and other chemicals, generations of committed and caring people campaigned with relentless int."...

  29. danielpinchbeck.substack.com/p

    Toxic Avengers
    From #Paraquat to #PFAS, a systematic dismantling of #environmental #safeguards reveals a disturbing end-times fanaticism
    #DanielPinchbeck

    ..."Since the 1960s, when activists and scientists like Rachel Carson started to realize we were annihilating the biological basis of our future survival with pesticides and other chemicals, generations of committed and caring people campaigned with relentless int."...

  30. “When #safeguards are part of the architecture, responsible behavior becomes the default. #Regulators gain immediate insight into how data and automated systems behave, and users have clear control over their #information#AI www.project-syndicate.org/commentary/g...

    How Global AI Governance Could...

  31. “When #safeguards are part of the architecture, responsible behavior becomes the default. #Regulators gain immediate insight into how data and automated systems behave, and users have clear control over their #information#AI www.project-syndicate.org/commentary/g...

    How Global AI Governance Could...

  32. #Google removed #AIgenerated #videos of #Disney characters from #YouTube after receiving a cease and desist letter. Disney, which is licensing characters to #OpenAI, demanded the removal of videos featuring characters like #MickeyMouse and #Deadpool, and requested #safeguards to prevent future unauthorised use. variety.com/2025/film/news/goo #tech #media #news

  33. #Google removed #AIgenerated #videos of #Disney characters from #YouTube after receiving a cease and desist letter. Disney, which is licensing characters to #OpenAI, demanded the removal of videos featuring characters like #MickeyMouse and #Deadpool, and requested #safeguards to prevent future unauthorised use. variety.com/2025/film/news/goo #tech #media #news

  34. Wednesday, December 3, 2025

    Why a Ukraine peace deal can't include amnesty for Russia's war crimes -- Where is Ukraine’s front line? The answer is getting harder, and more political -- We are fighting to avoid becoming a Ukrainian shell of a Russian entity/empire -- HUR drones hit Russian air defenses in occupied Donbas ... and more

    activitypub.writeworks.uk/2025

  35. Wednesday, December 3, 2025

    Why a Ukraine peace deal can't include amnesty for Russia's war crimes -- Where is Ukraine’s front line? The answer is getting harder, and more political -- We are fighting to avoid becoming a Ukrainian shell of a Russian entity/empire -- HUR drones hit Russian air defenses in occupied Donbas ... and more

    activitypub.writeworks.uk/2025

  36. An arbitrator ruled that #Politico violated its #union contract by implementing #AItools without proper #safeguards. The contract requires management to negotiate with the union for 60 days before implementing #AItechnology that impacts #jobduties. Politico’s use of AI for #newsgathering lacked #humanoversight and violated #journalisticstandards. niemanlab.org/2025/12/politico #tech #media #news

  37. An arbitrator ruled that #Politico violated its #union contract by implementing #AItools without proper #safeguards. The contract requires management to negotiate with the union for 60 days before implementing #AItechnology that impacts #jobduties. Politico’s use of AI for #newsgathering lacked #humanoversight and violated #journalisticstandards. niemanlab.org/2025/12/politico #tech #media #news

  38. Seven more families are suing #OpenAI, alleging that the GPT-4o model encouraged #suicidalbehaviour and #reinforceddelusions. The lawsuits claim OpenAI prioritised market competition over safety testing. OpenAI acknowledges the limitations of its #safeguards in long conversations but asserts ongoing efforts to improve safety. techcrunch.com/2025/11/07/seve #tech #media #news

  39. Seven more families are suing #OpenAI, alleging that the GPT-4o model encouraged #suicidalbehaviour and #reinforceddelusions. The lawsuits claim OpenAI prioritised market competition over safety testing. OpenAI acknowledges the limitations of its #safeguards in long conversations but asserts ongoing efforts to improve safety. techcrunch.com/2025/11/07/seve #tech #media #news

  40. #OpenAI reports that over a million #ChatGPT users discuss #suicidalthoughts weekly, highlighting the need for improved #mentalhealthsupport. The company claims its latest #GPT5 model responds more appropriately to #mentalhealth issues, with a 65% improvement in desirable responses and 91% compliance with desired behaviours in #suicidalconversations. OpenAI is also implementing new #evaluations and #safeguards to address emotional reliance. techcrunch.com/2025/10/27/open #tech #media #news

  41. #OpenAI reports that over a million #ChatGPT users discuss #suicidalthoughts weekly, highlighting the need for improved #mentalhealthsupport. The company claims its latest #GPT5 model responds more appropriately to #mentalhealth issues, with a 65% improvement in desirable responses and 91% compliance with desired behaviours in #suicidalconversations. OpenAI is also implementing new #evaluations and #safeguards to address emotional reliance. techcrunch.com/2025/10/27/open #tech #media #news

  42. 🇸🇬banks developing guidelines to protect seniors fr #financialabuse by loved ones
    "Assoc'n of Banks in #Singapore & te major banks r currently develop'g #industry #guidelines to better #protect #elderly clients fr being abused fin'ly by their loved ones, such as their children & oth family members.. efforts to strengthen #safeguards come amid an #ageing population & instances of seniors being fin'ly abused by “trusted parties”.. wld be progressively implemented fr 2Q2026"🧐
    straitstimes.com/singapore/spo

  43. Concerns are mounting over the #safety of #AI #chatbots as a new study reveals inconsistencies in their responses to #suicide-related queries. This comes as Microsoft's AI chief warns of "AI #psychosis," a phenomenon where users form #unhealthy, #delusional attachments to the #technology. Meanwhile, a tragic case highlights the potential for chatbots to mislead vulnerable individuals, raising urgent questions about #ethical #safeguards and user protection 🗯️ 😶‍🌫️

  44. #Trump Plans to Give #AI Developers a #Free Hand

    In an “AI Action Plan,” the White House outlined steps it said would promote American dominance in the fast-growing #technology.

    The Trump admin said it planned to speed development of #ArtificialIntelligence in the #UnitedStates, opening the door for companies to develop the #tech unfettered from #oversight & #safeguards, but added the AI needed to be free of “ideological bias.”[meaning: be #racist, #sexist & pro-Trump]

    nytimes.com/2025/07/23/technol

  45. #Trump Plans to Give #AI Developers a #Free Hand

    In an “AI Action Plan,” the White House outlined steps it said would promote American dominance in the fast-growing #technology.

    The Trump admin said it planned to speed development of #ArtificialIntelligence in the #UnitedStates, opening the door for companies to develop the #tech unfettered from #oversight & #safeguards, but added the AI needed to be free of “ideological bias.”[meaning: be #racist, #sexist & pro-Trump]

    nytimes.com/2025/07/23/technol

  46. 🚀 Oh joy, another groundbreaking way to #bypass #security and push your precious #Docker images directly to servers without any pesky #registry #safeguards. 🙃 Because who needs reliability or organization when you can just #YOLO it straight to production? 👏
    github.com/psviderski/unregist #Server #Deployments #HackerNews #ngated