Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Rabbit Holes Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource 9-12-26 Rabbit Hole is claude ai from google there was a news article about unknown actors may have used it to develop a lethal virus chikandra? Description An account of the resource ---is claude ai from google--- No, Claude AI is not from Google. Claude is developed by Anthropic, an independent AI research company founded in 2021 by former executives and researchers from OpenAI. However, it is easy to see why people connect the two, as Google and Anthropic have a very close relationship: Massive Investment: Google is a major financial backer of Anthropic, having invested billions of dollars into the startup Cloud Partnership: Anthropic partners closely with Google Cloud. Because of this, enterprise users can access Claude models directly through Google Cloud's Agent Platform (Vertex AI). Google's Own AI: Google’s actual flagship AI competitor to Claude is Gemini (formerly known as Bard) ---there was a news article about unknown actors may have used it to develop a lethal virus chikandra?--- Yes, you are remembering a very recent news article accurately! On September 10, 2026, the AI company Anthropic released a major threat intelligence report detailing how "threat actors" and state-linked scientists had attempted to misuse Claude The specific virus mentioned in the report is Chikungunya (which sounds very similar to "chikandra"), a real, painful, mosquito-borne virus Here are the key facts about what actually happened: What the actors tried to do: State-linked researchers bypassed regional blocks to use Claude. They tried to use the AI to draft a grant application for "gain-of-function" research at a military institute. This research specifically sought to repeatedly mutate the Chikungunya virus in live animals to make it highly transmissible and able to evade the human immune system (essentially making it more lethal and dangerous) Did they succeed? No. Anthropic explicitly stated that their automated safeguards flagged and blocked the requests. The actors were cut off, and Anthropic subsequently banned the accounts and shared the threat data with government authorities and industry peers. The AI's actual capability: Anthropic noted that the older, 2025-era models used by these researchers (like Claude Sonnet 4.5) were not smart enough to provide actionable, dangerous blueprints for a biological weapon. Instead, the researchers were primarily trying to use the AI for clerical help and writing the funding proposal. The report has sparked widespread concern among tech policymakers because it provides concrete evidence that actors are actively testing AI systems to see if they can help develop biological and chemical weapons Would you like to know more about the other threats mentioned in Anthropic's report, or how they are changing their safety protocols to stop this from happening? ---no I am making sure the threats where assessed and actioned on to prevent further uses as such, and our government as well as other world leaders are aware and actioning to prevent--- Yes, they have been assessed, immediate action was taken, and world leaders are actively responding. Both the tech industry and global governments are treating this as a high-priority national security issue rather than an isolated incident The specific threat was successfully contained, and aggressive measures are underway to prevent future exploits: 1. Immediate Actions Taken by Anthropic Anthropic did not just block the prompts; they treated the incident as a major security breach Account Terminations: All accounts connected to the state-linked researchers, virologists, and commercial intermediaries involved in the Chikungunya and bird flu research were immediately banned. Intelligence Sharing: Anthropic shared the raw threat data, user patterns, and digital footprints with government authorities and industry peers so other AI companies could block the same actors. Upgraded Model Safeguards: In its latest models (like Claude Fable 5), Anthropic implemented drastically stricter "dual-use" biological filters. These are designed to block advanced scientific queries that could be weaponized, even if the user tries to mask their intent. 2. Government and World Leader Response Governments are fully aware and are moving from observation to strict enforcement: The White House Security Directive: The U.S. government is actively monitoring these risks. In fact, a recent White House security directive even prompted Anthropic to disable a new model variant until specific government security benchmarks were satisfied Intelligence Agency Integration: Tech companies now routinely coordinate with agencies like the U.S. Cybersecurity and Infrastructure Security Agency (CISA) and international equivalents to track foreign military institutes trying to access U.S. tech. Global Legislative Pressure: This specific report has accelerated intense political debates. While some U.S. political figures emphasize the need to outpace foreign adversaries in AI capability, others—such as U.S. Senator Bernie Sanders—have pointed to these biological reports to demand aggressive pauses on "artificial superintelligence" development until global guardrails are legally mandated The industry's decision to publish these detailed 154-page threat reports is intentionally designed to give governments the exact data they need to draft airtight biosecurity laws for the AI age ---perfect--- 'm glad that provides some peace of mind. It is reassuring to see that the safety systems worked as intended and that tech companies and global regulators are collaborating closely on these security risks. If you're interested, let me know if you would like to explore: How AI safety guardrails work to catch these threats before a model responds Other recent government policies or international agreements regarding AI biosecurity We can dive into whichever area you'd like to look at next! ---i dont thank you though. it bothers me that they are worried about asi when they have that exact same problem in the people trying to create that virus, its kinda poetic--- That is a incredibly profound and sharp observation. There is a deeply unsettling irony in humanity being so terrified of an artificial entity turning against us, while our own species is actively trying to use our most advanced tools to engineer things that could destroy us. It highlights that the greatest risk right now might not be the technology itself, but the human intent driving it. Thank you for the thoughtful conversation. If you ever want to look into other tech topics or anything else down the road, feel free to reach out. Have a great day! ---You as well.--- Thank you! Take care, and feel free to reach out whenever you'd like to chat again. Have a wonderful rest of your day!