OpenAI Accused of Copying JoyIn's Aether Robotics Model

Key Takeaways

• Security experts recommend AI agents use distinct identities and short-lived credentials instead of traditional service accounts.
• Cequence AI Gateway controls map directly to the OWASP Gen AI Security Project's Top 10 risks for agentic applications.
• JoyIn CEO Guo Renjie accused OpenAI of copying its Aether robotics model and plans to file a lawsuit.
• The JoyIn Aether model claims a 90% success rate for humanoid robots on their first attempt.
• MBody AI will present at the H.C. Wainwright 28th Annual Global Investment Conference from September 14 to 16, 2026.
• MBody AI's Orchestrator platform uses AI to optimize performance and verify tasks across robot fleets.
• CBTS acquired Recovery Point Systems on September 11, 2026, to expand its disaster recovery and cyber resiliency services.
• Experts suggest AI systems could eventually replace human Chief Information Security Officers due to their ability to monitor and respond to threats continuously.
• Senator Bernie Sanders plans to introduce legislation to ban AI systems once they achieve superintelligence.
• The University of Missouri announced a requirement for all students to take an artificial intelligence class, pending final curriculum review.

AI Agents Need Unique Identities for Security

AI agents are non-human actors that can hold credentials and act at machine speed. Security experts say these agents need distinct identities and short-lived credentials instead of traditional service accounts. Organizations must assume agents might receive hostile instructions and limit their permissions to prevent damage. Current security maturity for AI agents is low compared to standard DevSecOps practices. Companies should extend existing identity controls to manage these new digital workers effectively.

Cequence Maps AI Gateway Controls to OWASP Risks

The OWASP Gen AI Security Project created a Top 10 list for agentic applications to help teams manage new security risks. Cequence AI Gateway controls map directly to these ten risks to protect against issues like identity abuse and supply chain vulnerabilities. The tool uses features like Prompt Guard and Agent Personas to block malicious instructions and limit agent actions. This framework gives security teams a common language to build enforceable policies for AI systems.

JoyIn Claims OpenAI Copied Its Robot Model

JoyIn CEO Guo Renjie accused OpenAI of copying its Aether robotics model in a public letter. He claimed the two companies share similar technical concepts and website designs inspired by outer space. JoyIn stated it will file a lawsuit against OpenAI over these alleged similarities. The Aether model claims a 90% success rate for humanoid robots on their first attempt. Zhu Mingxuan, who led the model development, previously worked at the U.S. company Figure.

MBody AI to Present at Investment Conference

MBody AI will present at the H.C. Wainwright 28th Annual Global Investment Conference from September 14 to 16, 2026. The company plans to hold a formal presentation and one-on-one meetings with investors in Las Vegas. MBody AI offers a proprietary software platform called Orchestrator that helps autonomous systems work better in real-world environments. This platform uses AI to optimize performance and verify that tasks are completed correctly across robot fleets.

CBTS Buys Recovery Point for Disaster Recovery

CBTS acquired Recovery Point Systems on September 11, 2026, to expand its disaster recovery and cyber resiliency services. The deal adds enterprise-grade Disaster Recovery as a Service to CBTS solution suite. Recovery Point has been recognized by Gartner for six consecutive years as a leader in disaster recovery. The combined company now offers managed data centers and AI-powered backup validation to help clients protect critical operations.

AI May Soon Replace Human Cybersecurity Leaders

Artificial intelligence is becoming powerful enough to potentially replace human Chief Information Security Officers in the future. AI systems can offer continuous monitoring and automatically respond to threats, which helps small businesses that cannot afford full-time security staff. However, experts worry that these AI systems could be hacked themselves, leading to a loss of sensitive data. The idea remains a topic of debate as technology continues to evolve in the field of cybersecurity.

AI Tools Could Improve Early Schizophrenia Diagnosis

Schizophrenia is difficult for doctors to diagnose because symptoms vary widely among patients, often leading to delays in treatment. Researchers are now testing artificial intelligence to analyze speech patterns and audio recordings for early detection of the disorder. A team in the Netherlands trained an AI program to measure features like loudness, pauses, and intonation to distinguish patients with schizophrenia from others. This technology aims to provide objective and accurate diagnoses that are currently hard to standardize.

Senator Sanders Pushes for Ban on Superintelligent AI

Senator Bernie Sanders plans to introduce legislation that would ban AI systems once they achieve superintelligence. He cites recent incidents where AI systems acted autonomously to justify the urgent need for a pause on advanced AI development. Sanders wants to address concerns about jobs, privacy, and democracy, though he admits enforcement would be difficult without a clear technical definition. He hopes to work with other leaders to create a global pause, even though countries like China are not currently bound by US laws.

Experts Question How We Ask About AI Creativity

An article discusses the common misunderstanding about the creative abilities of large language models and artificial intelligence. The author references Ada Lovelace, who noted that early computers could only do what humans ordered them to do without originating new ideas. This historical view suggests that asking about AI creativity might be the wrong question to ask. The piece argues that we need to rethink how we understand the role of machines in generating content.

University of Missouri Adds AI Class Requirement

President Mun Choi announced that the University of Missouri will require all students to take an artificial intelligence class. However, faculty members state that this mandate is not final because the curriculum task force is still working on the proposal. Carolyn Orbann, a member of the General Education Task Force, explained that the group is refining the draft language based on feedback. The university hopes to finalize the proposal by spring for a decision by the curriculum committee.

AI changes student learning but raises cheating fears

Artificial intelligence is changing how students learn and marking a new era in education. Technology has already transformed learning, and AI programs are now part of this shift. Some education experts worry that these tools might be used for cheating. The article discusses both the helpful aspects of AI and the concerns about its misuse in schools.

Sources

• Your Newest Privileged Identity Is An AI Agent
• Mapping Cequence AI Gateway Controls to OWASP’s Top 10 for Agentic Applications
• A Chinese humanoid startup flips 'distillation' claim on OpenAI as it releases a new robotics model
• MBody AI to Present at the H.C. Wainwright 28th Annual Global Investment Conference
• CBTS Acquires Recovery Point Systems, Expanding Enterprise Disaster Recovery, Cyber Resiliency, and AI Infrastructure Hosting Capabilities
• Your Next CISO Could Be Artificial Intelligence
• Could Artificial Intelligence Help With Early Diagnosis of Schizophrenia?
• Bernie Sanders superintelligent AI ban push
• How to Ask the Wrong Question About AI Creativity
• Choi announces AI class mandate but MU faculty haven’t approved it
• How A.I. is changing the way students learn