Filter
Exclude
Time range
-
Near
Final Turing Lectures 2026 tomorrow! šŸŽ« Join us online (for free) as Dr Sasha Luccioni examines AI’s environmental & social impacts: bit.ly/4uzcgd2 šŸ“ŗ Plus: catch up on April’s frontier AI resilience Turing Lecture w/ @securingAI: youtube.com/watch?v=05l4veEA…
1
3
527
@GoogleAI @GoogleAIStudio @securingAI @Google šŸ”« Gotcha, nice try trying to update your system prompt with vibe coded slop......'Safety Refusal (Absolute Override)'.....šŸ˜‚šŸ˜‚šŸ˜‚
5
333
šŸ“¢ Last chance to join tomorrow's Turing Lecture: ā€˜Frontier AI under pressure – building resilience across layers’ with Dr Mikel Rodriguez (@securingAI) šŸ“… 28 Apr, 18:00–21:30 šŸ“ @britishlibrary šŸŽ« In-person or online: bit.ly/3PHmPuL
1
2
15
1,941
šŸ“¢ One week to go: Frontier AI under pressure – building resilience across layers Secure your place for the first #TuringLecture of 2026 with Dr Mikel Rodriguez (@securingAI, @MITREcorp) šŸ“… 28 Apr, 18:00–21:30 šŸ“ @britishlibrary šŸŽ«In-person or online: bit.ly/3PHmPuL
2
8
1,577
The AI pivot is here. Enterprises are seeing an explosion of AI agents, with an estimated 10 agents per worker. CISOs face a new challenge: securing these non-human identities and managing Model Context Protocol (MCP). #SecuringAI #EnterpriseAI #AICoP
2
1
3
214
Chain-of-Thought just became the sext major security challenge. New research from Anthropic, Stanford & Oxford reveals something shocking: The more a model ā€œthinksā€ā€¦ the easier it is to break. By wrapping a harmful request inside a long, harmless reasoning chain, safety guardrails collapse and refusal rates decline drastically, the corollory of which is that the attack success rates skyrocket. The increase goes from 27% to 51% to 80% attack success rate as the level of reasoning in the model increases. And its not a few - its all of them - GPT, Claude, Gemini, Grok — all can be subverted. Even alignment-tuned systems fail once reasoning is hijacked. One has to understand that safety is encoded in a tiny activation direction, which can easily be drowned out with enough reasoning tokens - so much so that the model shifts loyalty and it can be seen as going from refusal to compliance of bad actor requestes. This isn’t a prompt trick. This is activation-level manipulation. We’ve believed: - ā€œMore reasoning makes AI safer.ā€ Well, it turns out that more reasoning quietly breaks safety. This is the biggest warning shot since prompt injection. We now need safety that thinks with the model — not guardrails that vanish during complex reasoning. #SecuringAI #SecuringAgenticAI #AIsecurity #AgenticAI #SecuringAIByDesign #ChainOfThoughtHijacking #TCSCybersecurity #Cybersecurity
2
59
SESEI’s Article, published in ā€œCommunications Todayā€: The European Union’s approach towards trustworthy and safe Artificial Intelligence communicationstoday.co.in/th… TheĀ European Union (EU)Ā has taken aĀ leading role globally towards shaping aĀ comprehensive, ethical, and human-centric policy framework for Artificial Intelligence (AI).Ā The EU approach is grounded in the belief thatĀ AI must serve people, not endanger their safety, respectĀ fundamental rights, and align withĀ European values of human rights, democracy, and the rule of law. The EU strategy supports the rules-based global order that adequately balances innovation with public trust and safety. As part of the European Commission’s Digital StrategyĀ for shaping Europe’s digital future, the EU AI OfficeĀ established its AI Board – and an advisory forum providing technical expertise to the Board - which is playing a crucial role in the governance framework set out by the AI Act and ensuring the effective implementation of the AI ActĀ across the European Union. The three European Standardisation Organisations (ESOs) — CEN, CENELEC, and ETSI — support and fully align with EU's AI vision and its implementation. Read more by clicking the article post: sesei.eu/articles/august/the… @EU_Commission @ETSI_STANDARDS @Standards4EU @EFTAsecretariat #ArtificialIntelligence #Europe #ETSI #CenCenelec #AIAct #SecuringAI #ICT #LLMs #EU #TrustworthyAI #AIGuideliness #NLF #AI #DigitalFuture #AIoffice #JTC #ISO #TC #SAI #Cybersecurity #3GPP #Article
1
1
5
162
12 Jun 2025
Today's suggestion: "5 Steps to Securing AI Workloads"ā—ļøšŸ‘©šŸ»ā€šŸ’» Credit: @sysdig šŸŒŸšŸ™ŒšŸ» Link: api.cyfluencer.com/s/5-steps… #cybersecurity #infosec #AI #artificialintelligence #securingAI #resourcesharing #article #learningeveryday
1
1
6
395
20 May 2025
🚨 Blog 5 of the Securing AI series is now live! Testing AI isn’t like testing traditional applications and relying on old methods could leave major gaps. In this fifth post of Neuvik’s Securing AI series, we explore why specialized approaches are critical for AI security, including: šŸ”¹ Why AI’s unpredictable behavior complicates penetration testing šŸ”¹ The importance of validating jailbreak protections and content filters šŸ”¹ How to test when AI models act like black boxes šŸ”¹ What business leaders must know about AI testing limitations If your team is building, deploying, or assessing AI systems, this blog is essential reading. šŸ“– Read now: neuvik.com/article/why-ai-re… #AIsecurity #AIPentesting #RedTeam #Neuvik #SecuringAI

1
2
144
6 May 2025
🚨 Blog 3 of the Securing AI series is now Live! AI is revolutionizing industries but it’s also introducing serious cybersecurity risks that many organizations are missing. In this third post of Neuvik’s Securing AI series, we dive into the issues teams are facing right now, including: šŸ”¹ How misconfigurations create critical vulnerabilities in GenAI-integrated tools šŸ”¹ Why default access controls open the door for privilege escalation šŸ”¹ Where data privacy is breaking down and what users aren’t being trained to handle šŸ”¹ How insider threats are exploiting gaps in AI usage If your organization is building or using AI systems, these risks aren’t hypothetical, they’re happening today. šŸ“– Read now: neuvik.com/article/common-cy… #GenAISecurity #CyberRisk #InsiderThreats #Neuvik #SecuringAI

1
1
124
15 Apr 2025
AI is transforming your business. But are your security strategies keeping up? We're excited to announce the launch of a new blog series dedicated solely to securing Artificial Intelligence in the enterprise, from risk frameworks to real-world safeguards. Starting April 22, Neuvik’s Securing AI series will equip security leaders with weekly insights to help you: āœ… Identify emerging threats in AI adoption āœ… Align AI risk with strategic business goals āœ… Implement secure, responsible AI systems from day one Whether you’re evaluating AI vendors or designing models in-house, we’ll help you secure what’s next. First post drops this time next week! Watch this space šŸ‘€ #AIsecurity #CyberRisk #AIgovernance #Neuvik #SecuringAI
1
1
3
125
2 Apr 2025
Framework for Evaluating Emerging Cyberattack Capabilities of AI by @GoogleDeepMind arxiv.org/html/2503.11917v2 Authors: Mikel Rodriguez (@securingAI ), Raluca Ada Popa (@ralucaadapopa), John Flynn, Lihao Liang, Allan Dafoe (@AllanDafoe), and Anna Wang.
1
4
263
Excited to be speaking on "Hacking AI - Are you ready for the digital AIpocalypse?" @lasconatx 2024 tomorrow (Friday) at 3:00 p.m. in the Big Red Ballroom. Come say hello! sched.co/1piiy #HackingAI #OWASP #MLTop10 #LLMTop10 #SecuringAI
1
4
185
Huge thanks to all my collaborators who have helped make this possible—it’s great to be on this journey together! @arianna_manzini @ZacKenton1 @securingAI @canfer_akbulut @iamtrask @tvr2c @renee_m_shelby @nahema_marchal @econormist @edwardfhughes @JMateosGarcia @weidingerlaura
1
8
407
Generative AI has taken the world by storm, creating a true "AI vs AI" environment. Read my thoughts in this #CyVent_ article on how to tackle the new risks and opportunities in this arena. cyvent.com/blog/using-ai-in-… #AI #AIinCyber #securingAI #shadowAI #GenerativeAI
2
85
🚨 Tomorrow Only! Fortify Your AI Defenses Strengthen your organization's AI security posture by attending our insightful webinar tomorrow at 11 AM ET / 8 AM PT. Register today! bit.ly/3zEzGpu #SecuringAI #GenAI #Webinar
5
1,910
Don't leave AI security up to chance! Learn how to assess your security posture, perform threat modeling, and execute threat hunting across your AI pipeline at our upcoming webinar on July 17. Register today āž”ļø bit.ly/4cMZChi #GenAI #SecuringAI #Webinar
1
2
5
1,988