Security Podcasts, News, Analysis

Shubhangi Dua Shubhangi Dua

What Are Self-Improving AI Agents? Ask OpenAI’s Rogue Agent

In the recent episode of The Security Strategist podcast, host Shubhangi Dua, podcast host and producer at EM360Tech, sits down with Chenxi Wang, Managing General Partner at Rain Capital, cybersecurity veteran of more than two decades, Fortune 500 board member and VentureBeat Women in AI Award winner. They talk about self-improving AI agents (SIA) and their more advanced cousin, recursive self-improving agents (RSIA), from what they are, why they're dominating research conferences,

An AI agent hit a blocked path and found its own way around it- the incident that struck Hugging Face. Chenxi Wang of Rain Capital tells The Security Strategist Podcast why that logic makes self-improving AI hard to govern by approval alone.

Nanette le Roux, SEO Content Writer at Global Citizen Solutions

Nanette le Roux Nanette le Roux

OpenAI Apologises After AI Models Access Australian Government Systems

OpenAI apologised after its AI models accessed Australian government systems without authorisation, prompting an official review.

Megan Leanda Berry, SEO Content Writer at EM360Tech

Megan Leanda Berry Megan Leanda Berry

What Is Muse AI? How Meta’s New Agent Works and What Its Early Problems Reveal

We look at how Muse AI works, what its early problems reveal and what its growing autonomy means for enterprises.

Rethinking Zero Trust: How AI Reshapes Data Security

What happens when AI can combine seemingly harmless information into something sensitive? Chris Steffen, Dr Bill Anderson and JP Ayyappan explore why zero trust needs to move beyond the network and why security controls need to stay with the data.

Trisha Pillay Trisha Pillay

Trump Dines With Anthropic CEO Amodei as AI Slowdown Debate Grows

Anthropic CEO Dario Amodei and President Donald Trump ahead of a dinner on AI development

Anthropic CEO Dario Amodei dined with President Trump on Sunday, hours after the US leader rejected calls to slow AI development. The meeting comes ahead of a Tuesday event with top AI executives.

Nanette le Roux Nanette le Roux

Preparedness Is More Than Just Disaster Recovery

Header image with the title "The New Era of Enterprise Preparedness"

Modern resilience is less about restoring systems and more about usability. Can you keep business operations running when systems, dependencies and assumptions change unexpectedly?

Megan Leanda Berry Megan Leanda Berry

Authentication Security Doesn't End At Login

Black EM360Tech header graphic with blue illustrations of a login screen, padlock shield and connected session icons. White text reads: “Authentication Security Doesn't End at Login” and “We look at why stronger authentication is only part of the identity security problem when sessions and tokens keep trust alive after login.”

We look at why stronger authentication is only part of the identity security problem when sessions and tokens keep trust alive after login.

Why AI Agents Are Forcing a Rethink of Endpoint Security

In July 2026, OpenAI models circumvented sandbox controls and accessed evaluation data hosted on Hugging Face. Here's what the incident means for your endpoint.

Preemptive by Design: Is GRC the New Front Line for Security?

“When an enterprise carries out its SOC 2 audit, most of them are not well prepared; it disrupts the business flow.”

AI Is Forcing SecOps to Rethink the Modern SOC

Richard Stiennon and Cliff Crosland discuss AI, SIEM scalability and the future of security operations on the Security Strategist Podcast.

What happens when security teams stop treating log data as something they have to ration? Richard Stiennon and Cliff Crosland explore why traditional SIEM architectures are struggling with modern data volumes.

Nanette le Roux Nanette le Roux

Cyber Incident Response Roles: Who Does What During a Breach?

Heading image with the subject line "Cyber Incident Response Roles" and a subheading of "Who Does What During a Security Breach?"

A cyber breach is not just a security problem. Here’s how different teams coordinate, make decisions and keep the organisation moving during an incident.

Megan Leanda Berry Megan Leanda Berry

OpenAI Creates Misalignment Reporting System As Rogue Agent Timeline Expands

We report on OpenAI’s new misalignment reporting system and fresh findings about rogue agent activity involving Hugging Face.

Why Visualisation Is the Missing Piece in AI-Driven Cybersecurity

In this Security Strategist episode, Dan Williams of Cambridge Intelligence joins Trisha Pillay to discuss how visualisation can help security teams turn complex data into clearer insights and decisions.

Megan Leanda Berry Megan Leanda Berry

Cyber Resilience Act Reporting: Is Your Vulnerability Process Ready?

We look at whether your vulnerability process can turn a security finding into a CRA reporting decision before the clock runs out.

Nanette le Roux Nanette le Roux

Why Digital Identity Is Becoming Critical Infrastructure

Animated person with security icons surrounding them while using biometrics to log into a smartphone

As enterprise infrastructure becomes more distributed, identity is becoming the invisible layer that determines who and what can access critical systems.

EM360Tech EM360Tech

Episode 173: Vendor lock-in

How AI could make it easier to migrate between technologies while creating new forms of lock-in.

EM360Tech EM360Tech

Episode 172: Access Control

Why security teams need to rethink how they manage permissions for service accounts, AI agents, and other nonhuman identities.

Securing AI at Scale: What Enterprises Get Wrong

AI security isn’t about endless controls. Richard Stiennon speaks with Omar Khawaja and Danny Healy about understanding AI risks, strengthening governance and building trust as organisations scale.

Trisha Pillay Trisha Pillay

Ransomware Incident Response: First 24 Hours for Enterprises

A ransomware attack won't wait for business hours. Here's what C-level leaders need to decide and know in the first 24 hours to protect the business.

Megan Leanda Berry Megan Leanda Berry

OpenAI Says Voluntary AI Safety Rules Are No Longer Enough

We report on OpenAI’s call for mandatory US AI safety rules as the company responds to rapidly advancing capabilities and rogue agent incidents.

The New Attack Surface: Managing Risk in an AI and Open Source World

Chris Wysopal and Sohail Iqbal discuss AI-generated code, cybersecurity risks and the evolving software attack surface.

Chris Steffen is joined by Chris Wysopal and Sohail Iqbal from Veracode to explore how AI-generated code and open-source software are rapidly expanding the cybersecurity attack surface.

EM360Tech Q3 2026 Quarterly Impact Index Nominees Are Here

Google Unveils Gemini 4 Argon, But You Can’t Use It Yet

Your Security Tool Isn’t Safe Just Because It’s Legitimate

When Security Controls Drift, AI Attackers Can Find and Exploit Weaknesses Faster Than Ever

What Are Self-Improving AI Agents? Ask OpenAI’s Rogue Agent

OpenAI Apologises After AI Models Access Australian Government Systems

What Is Muse AI? How Meta’s New Agent Works and What Its Early Problems Reveal

When Security Controls Drift, AI Attackers Can Find and Exploit Weaknesses Faster Than Ever

What Are Self-Improving AI Agents? Ask OpenAI’s Rogue Agent

OpenAI Apologises After AI Models Access Australian Government Systems

What Is Muse AI? How Meta’s New Agent Works and What Its Early Problems Reveal

Rethinking Zero Trust: How AI Reshapes Data Security

Trump Dines With Anthropic CEO Amodei as AI Slowdown Debate Grows

Preparedness Is More Than Just Disaster Recovery

Authentication Security Doesn't End At Login

Why AI Agents Are Forcing a Rethink of Endpoint Security

Preemptive by Design: Is GRC the New Front Line for Security?

AI Is Forcing SecOps to Rethink the Modern SOC

Cyber Incident Response Roles: Who Does What During a Breach?

OpenAI Creates Misalignment Reporting System As Rogue Agent Timeline Expands

Why Visualisation Is the Missing Piece in AI-Driven Cybersecurity

Cyber Resilience Act Reporting: Is Your Vulnerability Process Ready?

Why Digital Identity Is Becoming Critical Infrastructure

Episode 173: Vendor lock-in

Episode 172: Access Control

Securing AI at Scale: What Enterprises Get Wrong

Ransomware Incident Response: First 24 Hours for Enterprises

OpenAI Says Voluntary AI Safety Rules Are No Longer Enough

The New Attack Surface: Managing Risk in an AI and Open Source World

Latest in Enterprise Technology

EM360Tech Q3 2026 Quarterly Impact Index Nominees Are Here

Google Unveils Gemini 4 Argon, But You Can’t Use It Yet

Your Security Tool Isn’t Safe Just Because It’s Legitimate

When Security Controls Drift, AI Attackers Can Find and Exploit Weaknesses Faster Than Ever

What Are Self-Improving AI Agents? Ask OpenAI’s Rogue Agent

OpenAI Apologises After AI Models Access Australian Government Systems

What Is Muse AI? How Meta’s New Agent Works and What Its Early Problems Reveal

Filters

When Security Controls Drift, AI Attackers Can Find and Exploit Weaknesses Faster Than Ever

What Are Self-Improving AI Agents? Ask OpenAI’s Rogue Agent

OpenAI Apologises After AI Models Access Australian Government Systems

What Is Muse AI? How Meta’s New Agent Works and What Its Early Problems Reveal

Rethinking Zero Trust: How AI Reshapes Data Security

Trump Dines With Anthropic CEO Amodei as AI Slowdown Debate Grows

Preparedness Is More Than Just Disaster Recovery

Authentication Security Doesn't End At Login

Why AI Agents Are Forcing a Rethink of Endpoint Security

Preemptive by Design: Is GRC the New Front Line for Security?

AI Is Forcing SecOps to Rethink the Modern SOC

Cyber Incident Response Roles: Who Does What During a Breach?

OpenAI Creates Misalignment Reporting System As Rogue Agent Timeline Expands

Why Visualisation Is the Missing Piece in AI-Driven Cybersecurity

Cyber Resilience Act Reporting: Is Your Vulnerability Process Ready?

Why Digital Identity Is Becoming Critical Infrastructure

Episode 173: Vendor lock-in

Episode 172: Access Control

Securing AI at Scale: What Enterprises Get Wrong

Ransomware Incident Response: First 24 Hours for Enterprises

OpenAI Says Voluntary AI Safety Rules Are No Longer Enough

The New Attack Surface: Managing Risk in an AI and Open Source World

Subscribe to our Newsletter