【US Patent】Defending Against Attacks from "Claude Mythos" Class Ultra-High Performance AI Models! Patent Technology Related to Authenticity Verification "Authenticity Verification System®", which has Passed "PCT International Patent Examination (157 Member Countries)", Successfully Granted US Patent Rights
Cycal Trust has successfully secured a US patent for its core 'Authenticity Verification System®' technology. This establishes a protection base for its technology in the world's largest markets, enabling defense against sophisticated AI-generated misinformation and ensuring the authenticity of various assets.
📋 Article Processing Timeline
- 📰 Published: May 5, 2026 at 03:46
- 🔍 Collected: May 4, 2026 at 19:01
- 🤖 AI Analyzed: May 4, 2026 at 19:14 (12 min after Collected)
Cycal Trust Inc. (Headquarters: Shibuya-ku, Tokyo; Representative Director: Takeshi Sue; hereinafter "Cycal Trust") announces the successful acquisition of patent rights in the "United States" for its PCT application (international application under the Patent Cooperation Treaty) corresponding to Japanese Patent "No. 7371301". The US patent number will be announced separately after its assignment.
This patent is one of multiple patented technologies comprising Cycal Trust's core technology, the "Authenticity Verification System®", and relates to elemental technology for ensuring the authenticity (trust) of all assets (including "physical assets" such as goods and products, "non-physical assets" such as data and content, and "hybrid assets" that integrate both, such as RWAs and digital twins). With the successful acquisition of rights in "the United States" following "Japan", Cycal Trust has established a protection foundation for this technology portfolio in the world's largest semiconductor, financial, critical infrastructure, medical, and government/defense markets, among others.
Taking this as an opportunity, Cycal Trust intends to fully commence offering the "Authenticity Verification System®" not only to the aforementioned individual industries but also to supply chain and traceability providers across these industries, as a defense against the ultra-high performance AI "Claude Mythos Preview" and other "Claude Mythos-class AIs" expected to emerge continuously throughout the year.
Chapter 1: The Reality of Multiple "Claude Mythos-Class AIs"
Anthropic has indicated a policy of providing "Claude Mythos Preview" as "Project Glasswing" exclusively for the discovery and remediation of vulnerabilities in critical software. The UK NCSC (National Cyber Security Centre) has pointed out that frontier AI models have changed the cost structure for both attackers and defenders in areas such as "zero-day (undisclosed vulnerability) discovery", cryptographic challenges, and "exploit code creation". The Australian ASD (Signals Directorate) has also warned that the premise of considering capabilities like "Claude Mythos Preview" and "OpenAI GPT-5.5" as "rare or unique" is no longer valid.
Multiple cybersecurity companies anticipate that "advanced AIs with higher-level cyber capabilities will become generalized" in 2026. The threat to be addressed is not a single specific model, but a structural threat where multiple "Claude Mythos-class AIs" coexist.
Chapter 2: The Core Risk Is Not "Intrusion" but "Adoption of Misinformation"?
The core threat of "Claude Mythos-class AI" is not just the acceleration of attacks, but the ability to make AI-generated misinformation be adopted as genuine by both humans and existing systems.
Fake remittance instructions, identity verification, SBOMs (Software Bill of Materials), inspection certificates, audit logs, vulnerability remediation reports, AI analysis results—all flow into business systems with a "seemingly genuine" finish. Conventional intrusion prevention, malware detection, communication monitoring, and other vulnerability remediation alone cannot stop these.
Chapter 3: The Solution Is Not "More Advanced AI" but a "Chain of Authenticity (Trust)"?
The idea of detecting AI-generated misinformation with AI leads to a performance race between attacking AI and defending AI. What is needed is a mechanism to determine whether AI output can be connected to a chain of authenticity (trust), rather than simply believing the AI output.
Cycal Trust's "Authenticity Verification System®" integrates multiple patented technologies to "authenticate and verify" transaction information, supply chain information, AI output, and other verification results for all assets (physical assets, non-physical assets, and hybrid assets) from the perspective of authenticity (trust). The operating principle is clear: "Do not adopt information whose authenticity (trust) cannot be verified."
Chapter 4: What is the Difference Between Existing "Authenticity (Trust) Technologies" and the "Authenticity Verification System®"?
Existing technologies that ensure authenticity (trust) include "C2PA" for media provenance, "W3C (Verifiable Credentials)" for digital credentials, "Sigstore" for software signing, and "PKI (Public Key Infrastructure)"-based electronic signatures.
While these are effective in specific domains, in the era of "Claude Mythos-class AI", they commonly face limitations in three areas:
(1) Verification of authenticity for data whose correct answer is not fixed at the time of issuance, such as AI output.
(2) Detection when the signature key itself is forged or leaked.
(3) Seamless application to hybrid domains spanning physical and non-physical assets.
The "Authenticity Verification System®" is designed as an authenticity (trust) defense layer that covers all three of these areas.
This patent is one of multiple patented technologies comprising Cycal Trust's core technology, the "Authenticity Verification System®", and relates to elemental technology for ensuring the authenticity (trust) of all assets (including "physical assets" such as goods and products, "non-physical assets" such as data and content, and "hybrid assets" that integrate both, such as RWAs and digital twins). With the successful acquisition of rights in "the United States" following "Japan", Cycal Trust has established a protection foundation for this technology portfolio in the world's largest semiconductor, financial, critical infrastructure, medical, and government/defense markets, among others.
Taking this as an opportunity, Cycal Trust intends to fully commence offering the "Authenticity Verification System®" not only to the aforementioned individual industries but also to supply chain and traceability providers across these industries, as a defense against the ultra-high performance AI "Claude Mythos Preview" and other "Claude Mythos-class AIs" expected to emerge continuously throughout the year.
Chapter 1: The Reality of Multiple "Claude Mythos-Class AIs"
Anthropic has indicated a policy of providing "Claude Mythos Preview" as "Project Glasswing" exclusively for the discovery and remediation of vulnerabilities in critical software. The UK NCSC (National Cyber Security Centre) has pointed out that frontier AI models have changed the cost structure for both attackers and defenders in areas such as "zero-day (undisclosed vulnerability) discovery", cryptographic challenges, and "exploit code creation". The Australian ASD (Signals Directorate) has also warned that the premise of considering capabilities like "Claude Mythos Preview" and "OpenAI GPT-5.5" as "rare or unique" is no longer valid.
Multiple cybersecurity companies anticipate that "advanced AIs with higher-level cyber capabilities will become generalized" in 2026. The threat to be addressed is not a single specific model, but a structural threat where multiple "Claude Mythos-class AIs" coexist.
Chapter 2: The Core Risk Is Not "Intrusion" but "Adoption of Misinformation"?
The core threat of "Claude Mythos-class AI" is not just the acceleration of attacks, but the ability to make AI-generated misinformation be adopted as genuine by both humans and existing systems.
Fake remittance instructions, identity verification, SBOMs (Software Bill of Materials), inspection certificates, audit logs, vulnerability remediation reports, AI analysis results—all flow into business systems with a "seemingly genuine" finish. Conventional intrusion prevention, malware detection, communication monitoring, and other vulnerability remediation alone cannot stop these.
Chapter 3: The Solution Is Not "More Advanced AI" but a "Chain of Authenticity (Trust)"?
The idea of detecting AI-generated misinformation with AI leads to a performance race between attacking AI and defending AI. What is needed is a mechanism to determine whether AI output can be connected to a chain of authenticity (trust), rather than simply believing the AI output.
Cycal Trust's "Authenticity Verification System®" integrates multiple patented technologies to "authenticate and verify" transaction information, supply chain information, AI output, and other verification results for all assets (physical assets, non-physical assets, and hybrid assets) from the perspective of authenticity (trust). The operating principle is clear: "Do not adopt information whose authenticity (trust) cannot be verified."
Chapter 4: What is the Difference Between Existing "Authenticity (Trust) Technologies" and the "Authenticity Verification System®"?
Existing technologies that ensure authenticity (trust) include "C2PA" for media provenance, "W3C (Verifiable Credentials)" for digital credentials, "Sigstore" for software signing, and "PKI (Public Key Infrastructure)"-based electronic signatures.
While these are effective in specific domains, in the era of "Claude Mythos-class AI", they commonly face limitations in three areas:
(1) Verification of authenticity for data whose correct answer is not fixed at the time of issuance, such as AI output.
(2) Detection when the signature key itself is forged or leaked.
(3) Seamless application to hybrid domains spanning physical and non-physical assets.
The "Authenticity Verification System®" is designed as an authenticity (trust) defense layer that covers all three of these areas.