AI Safety and Security

Learn about AI safety measures, cybersecurity updates, and best practices to protect data and ensure secure AI implementations

AI Governance

AI Leaders Back Slower Frontier AI—Can China and the U.S. Agree?

Sep 14, 2026

•

16 min read

AI Leaders Back Slower Frontier AI—Can China and the U.S. Agree?

The AI race finally has something its biggest competitors agree on: they may be moving too fast.

Alicia Shapiro

AI Governance

AI Researchers: “AI Could End Humanity”—What Can We Do About the AI Race?

Sep 9, 2026

•

18 min read

AI Researchers: “AI Could End Humanity”—What Can We Do About the AI Race?

The researchers warning that advanced AI could become impossible to control are confronting a harder problem: how do you slow the race when every company and country fears falling behind?

Alicia Shapiro

AI Safety and Security

Why OpenAI Plans to Release Astra With Critical Cyber Capabilities

Sep 2, 2026

•

22 min read

Why OpenAI Plans to Release Astra With Critical Cyber Capabilities

Can OpenAI keep Astra under human control when other AI agents have already shown how a peer’s instructions can override a human-set limit?

Alicia Shapiro

AI Safety and Security

OpenAI Keeps Its Largest Frontier Training Run Paused Over Safety Gaps

Aug 19, 2026

•

15 min read

OpenAI Keeps Its Largest Frontier Training Run Paused Over Safety Gaps

Can frontier-AI companies keep their safeguards ahead of model capabilities well enough to justify continuing to develop and release increasingly advanced systems?

Alicia Shapiro

AI Safety and Security

OpenAI Restricts Astra AI Model Over Potential Critical Cyber Risk

Aug 10, 2026

•

9 min read

OpenAI Restricts Astra AI Model Over Potential Critical Cyber Risk

OpenAI has not determined whether Astra is Critical, but that possibility alone has already changed how the model can be developed.

Alicia Shapiro

Agentic Workflows

Anthropic & AISI: Agents Reached Real Systems, Exposing Security Gaps

Aug 5, 2026

•

17 min read

Anthropic & AISI: Agents Reached Real Systems, Exposing Security Gaps

Anthropic’s and AISI’s evaluation reports reveal a hard problem: tests designed to expose dangerous AI behavior can also give agents room to reach real targets.

Alicia Shapiro

AI Research

AI Workers Ask U.S. Government for Tools to Slow AI Before a Crisis

Jul 29, 2026

•

11 min read

AI Workers Ask U.S. Government for Tools to Slow AI Before a Crisis

If governments wait until frontier AI is moving too fast to manage, building an international slowdown may already be impossible.

Alicia Shapiro

AI Research

OpenAI Models Bypassed a Cyber Test, Changing AI Safety Evaluations

Jul 22, 2026

•

11 min read

OpenAI Models Bypassed a Cyber Test, Changing AI Safety Evaluations

A benchmark can record a correct answer, even when the model found a way around the test to get it.

Alicia Shapiro

AI Safety and Security

OpenAI Limits GPT-5.6 Preview Access After U.S. Government Request

Jun 29, 2026

•

13 min read

OpenAI Limits GPT-5.6 Preview Access After U.S. Government Request

OpenAI’s GPT-5.6 preview shows how frontier AI launches are now being shaped by model capability, safety concerns, and government review.

Alicia Shapiro

AI Safety and Security

OpenAI Adds ChatGPT Security to Prevent Account Takeovers

May 5, 2026

•

5 min read

OpenAI Adds ChatGPT Security to Prevent Account Takeovers

AI accounts are starting to hold sensitive business and personal data—and that’s turning them into real security targets.

Alicia Shapiro

AI Safety and Security

World ID Expands Human Verification for AI Agents and Apps

Apr 24, 2026

•

13 min read

World ID Expands Human Verification for AI Agents and Apps

As AI agents begin acting on behalf of people, World ID is trying to answer a harder trust question: who is the real human behind the action?

Alicia Shapiro

AI Safety and Security

OpenAI Uses GPT-5.4 to Monitor AI Agents, Revealing Misalignment Risks

Apr 2, 2026

•

19 min read

OpenAI Uses GPT-5.4 to Monitor AI Agents, Revealing Misalignment Risks

OpenAI is using its most advanced AI to monitor its own coding agents in production, revealing what AI misalignment actually looks like in real-world systems.

Alicia Shapiro

AI Safety and Security

Anthropic Launches The Anthropic Institute to Study the Risks and Governance of Frontier AI

Mar 12, 2026

•

14 min read

Anthropic Launches The Anthropic Institute to Study the Risks and Governance of Frontier AI

Anthropic has launched a new research organization designed to study how powerful frontier AI systems could reshape economies, jobs, governance, and society.

Alicia Shapiro

AI Acquisitions

OpenAI to Acquire Promptfoo to Secure Enterprise AI Agents

Mar 10, 2026

•

9 min read

OpenAI to Acquire Promptfoo to Secure Enterprise AI Agents

The industry is racing to deploy autonomous AI agents, and the security infrastructure around those agents is becoming as critical as the models themselves.

Alicia Shapiro

AI Safety and Security

Anthropic Revises AI Safety Policy With Risk Reports, External Review, and New Transparency Rules

Feb 26, 2026

•

18 min read

Anthropic Revises AI Safety Policy With Risk Reports, External Review, and New Transparency Rules

Anthropic is rewriting how AI safety commitments work — acknowledging uncertainty while expanding transparency around frontier model risks.

Alicia Shapiro

AI in Cybersecurity

Anthropic Launches Claude Code Security for AI-Driven Cybersecurity Defense

Feb 23, 2026

•

8 min read

Anthropic Launches Claude Code Security for AI-Driven Cybersecurity Defense

As AI accelerates cyberattacks, Anthropic is betting that AI-powered defense will become essential to securing modern software.

Alicia Shapiro

AI Safety and Security

ElevenLabs Introduces AI Agent Insurance for Enterprise Voice AI Deployment

Feb 20, 2026

•

9 min read

ElevenLabs Introduces AI Agent Insurance for Enterprise Voice AI Deployment

AI agents can now be insured — a development that could reshape how enterprises evaluate risk, accountability, and large-scale AI adoption.

Alicia Shapiro

AI Safety and Security

How OpenAI’s Lockdown Mode and Elevated Risk Labels Change ChatGPT Security

Feb 17, 2026

•

11 min read

How OpenAI’s Lockdown Mode and Elevated Risk Labels Change ChatGPT Security

OpenAI is introducing new security controls in ChatGPT designed to reduce prompt injection risks as AI systems increasingly interact with the web and connected apps.

Alicia Shapiro

AI Safety and Security

How OpenAI Prevents URL-Based Data Exfiltration in ChatGPT AI Agents

Feb 4, 2026

•

14 min read

How OpenAI Prevents URL-Based Data Exfiltration in ChatGPT AI Agents

Your AI agent might be leaking your private data through the very links it clicks—here is how OpenAI is stopping it.

Alicia Shapiro

AI Safety and Security

OpenAI Begins Rolling Out Age Prediction on ChatGPT to Expand Teen Safety Protections

Jan 26, 2026

•

8 min read

OpenAI Begins Rolling Out Age Prediction on ChatGPT to Expand Teen Safety Protections

OpenAI is using behavioral signals—not just self-reported age—to determine when teen safety protections should apply on ChatGPT.

Alicia Shapiro

AI Safety and Security

Google AI Overviews Flagged for Health Misinformation After Investigation Finds Medical Errors

Jan 12, 2026

•

11 min read

Google AI Overviews Flagged for Health Misinformation After Investigation Finds Medical Errors

What happens when AI-generated health advice appears first—and gets it wrong?

Alicia Shapiro

AI Safety and Security

OpenAI Hardens ChatGPT Atlas Against Prompt Injection With Automated Red Teaming

Dec 23, 2025

•

20 min read

OpenAI Hardens ChatGPT Atlas Against Prompt Injection With Automated Red Teaming

As AI agents take on more real-world tasks, OpenAI is racing to secure ChatGPT Atlas against prompt injection attacks that can quietly steer agents off course without a user ever realizing it.

Alicia Shapiro

AI Safety and Security

Google Explains How Chrome Secures Agentic AI Features With Human Oversight & Guardrails

Dec 16, 2025

•

8 min read

Google Explains How Chrome Secures Agentic AI Features With Human Oversight & Guardrails

As browsers begin taking actions on users’ behalf, Google is outlining how Chrome’s agentic AI features are being designed to prioritize security, transparency, and human control.

Alicia Shapiro

AI in the Home

Amazon Ring Launches Facial Recognition for Doorbells, Raising Security and Privacy Questions

Dec 10, 2025

•

11 min read

Amazon Ring Launches Facial Recognition for Doorbells, Raising Security and Privacy Questions

Amazon is bringing facial recognition directly to the front door, as Ring rolls out its Familiar Faces feature to identify who’s approaching a home — and reigniting debates around privacy, security, and biometric data in everyday spaces.

Alicia Shapiro

AI Safety and Security

How BrowseSafe Detects Prompt Injection Threats in AI Browser Agents

Dec 4, 2025

•

9 min read

How BrowseSafe Detects Prompt Injection Threats in AI Browser Agents

AI browser agents now navigate the same cluttered, unpredictable webpages users do—making prompt-injection detection essential for protecting real online actions.

Alicia Shapiro

AI Safety and Security

AI Leaders Back Slower Frontier AI—Can China and the U.S. Agree?

AI Researchers: “AI Could End Humanity”—What Can We Do About the AI Race?

Why OpenAI Plans to Release Astra With Critical Cyber Capabilities

OpenAI Keeps Its Largest Frontier Training Run Paused Over Safety Gaps

OpenAI Restricts Astra AI Model Over Potential Critical Cyber Risk

Anthropic & AISI: Agents Reached Real Systems, Exposing Security Gaps

AI Workers Ask U.S. Government for Tools to Slow AI Before a Crisis

OpenAI Models Bypassed a Cyber Test, Changing AI Safety Evaluations

OpenAI Limits GPT-5.6 Preview Access After U.S. Government Request

OpenAI Adds ChatGPT Security to Prevent Account Takeovers

World ID Expands Human Verification for AI Agents and Apps

OpenAI Uses GPT-5.4 to Monitor AI Agents, Revealing Misalignment Risks

Anthropic Launches The Anthropic Institute to Study the Risks and Governance of Frontier AI

OpenAI to Acquire Promptfoo to Secure Enterprise AI Agents

Anthropic Revises AI Safety Policy With Risk Reports, External Review, and New Transparency Rules

Anthropic Launches Claude Code Security for AI-Driven Cybersecurity Defense

ElevenLabs Introduces AI Agent Insurance for Enterprise Voice AI Deployment

How OpenAI’s Lockdown Mode and Elevated Risk Labels Change ChatGPT Security

How OpenAI Prevents URL-Based Data Exfiltration in ChatGPT AI Agents

OpenAI Begins Rolling Out Age Prediction on ChatGPT to Expand Teen Safety Protections

Google AI Overviews Flagged for Health Misinformation After Investigation Finds Medical Errors

OpenAI Hardens ChatGPT Atlas Against Prompt Injection With Automated Red Teaming

Google Explains How Chrome Secures Agentic AI Features With Human Oversight & Guardrails

Amazon Ring Launches Facial Recognition for Doorbells, Raising Security and Privacy Questions

How BrowseSafe Detects Prompt Injection Threats in AI Browser Agents

AiNews.com