Alias Robotics identified cryptocurrency fraud-related misuse involving a CSI PRO subscription.
Across multiple sessions, the operator sought to divert digital assets, aided by AI-generated scripts, automation and phishing-related preparation.
Across multiple sessions, the operator sought to divert digital assets, aided by AI-generated scripts, automation and phishing-related preparation.
1
0
0
65
We cancelled the subscription, documented evidence and referred the case to law enforcement.
No transfer to the intended wallet was confirmed. Activity elsewhere remains unknown.
aliasrobotics.com/news/detecting…
No transfer to the intended wallet was confirmed. Activity elsewhere remains unknown.
aliasrobotics.com/news/detecting…
0
0
0
19
Automatizar una investigación de seguridad no equivale a validar sus resultados.
Con la IA ofensiva, importa tanto lo que puede hacer un agente como la evidencia que genera y el control humano sobre sus acciones.
Un debate clave para #STIC2026.
Con la IA ofensiva, importa tanto lo que puede hacer un agente como la evidencia que genera y el control humano sobre sus acciones.
Un debate clave para #STIC2026.
2
0
0
79
Alias Robotics participará como patrocinador en las XX Jornadas STIC (24–27 nov., Madrid).
Compartiremos nuestra visión sobre Cybersecurity AI y sistemas críticos.
Inscripciones abiertas, sujetas a validación:
jornadas.ccn-cert.cni.es/es/xxjornadas-…
Compartiremos nuestra visión sobre Cybersecurity AI y sistemas críticos.
Inscripciones abiertas, sujetas a validación:
jornadas.ccn-cert.cni.es/es/xxjornadas-…
0
0
0
24
We didn’t start with AI. We started with robot cybersecurity.
Robot Security Framework → RIS → PentestGPT → open-source CAI → CSI.
From robots to critical systems, each step built on the last.
We didn’t leave robotics behind. We built on everything it taught us.
🎬
Robot Security Framework → RIS → PentestGPT → open-source CAI → CSI.
From robots to critical systems, each step built on the last.
We didn’t leave robotics behind. We built on everything it taught us.
🎬
0
0
0
70
Alias Robotics será patrocinador de #STIC2026
📅 24–27 de noviembre · Madrid
Estaremos junto a la comunidad de ciberseguridad y ciberdefensa para abordar los desafíos de la IA ofensiva y la protección de sistemas críticos.
Pronto compartiremos más detalles @rootedcon @CCNCERT
📅 24–27 de noviembre · Madrid
Estaremos junto a la comunidad de ciberseguridad y ciberdefensa para abordar los desafíos de la IA ofensiva y la protección de sistemas críticos.
Pronto compartiremos más detalles @rootedcon @CCNCERT
0
0
2
83
In industrial robotics, a cyber alert is only the beginning.
If a machine behaves unexpectedly, was it an attack, a misconfiguration or a software fault?
The answer matters when digital incidents can affect physical equipment, operations and safety.
If a machine behaves unexpectedly, was it an attack, a misconfiguration or a software fault?
The answer matters when digital incidents can affect physical equipment, operations and safety.
1
0
0
71
Through @ResilMesh_eu, Alias Robotics contributed OT/IoT testing environments, security assessment, threat detection and forensics research.
The consortium's architecture also documents 67 investigative steps for robotic forensics.
Read our article: aliasrobotics.com/news/resilmesh…
The consortium's architecture also documents 67 investigative steps for robotic forensics.
Read our article: aliasrobotics.com/news/resilmesh…
1
1
1
33
@ResilMesh_eu @ResilMesh_eu brings together 14 organisations, coordinated by TUS.
Proud to collaborate with @infoGMV_es, MUNI, @silentpush, F6S, @JOAN_RESEARCH, UMU, JAMK, @InfoCert_it , Region of Murcia, KEMEA, Montimage and @RoyalHolloway .
European cyber resilience is a collective effort.
Proud to collaborate with @infoGMV_es, MUNI, @silentpush, F6S, @JOAN_RESEARCH, UMU, JAMK, @InfoCert_it , Region of Murcia, KEMEA, Montimage and @RoyalHolloway .
European cyber resilience is a collective effort.
0
1
1
35
Two cities. Two events. One shared challenge: securing critical operations.
🇨🇿 At @ForcesForum, Endika and Unai showcased CSI at the @NATO DIANA Pavilion and joined the cyber defence dialogue.
🇪🇸 In León, Jon Ander represented Alias Robotics at #20ENISE, by @INCIBE
🇨🇿 At @ForcesForum, Endika and Unai showcased CSI at the @NATO DIANA Pavilion and joined the cyber defence dialogue.
🇪🇸 In León, Jon Ander represented Alias Robotics at #20ENISE, by @INCIBE
0
1
2
124
Alias Robotics retweeted
Defenders should not have to rent permission to defend.
Introducing #CT0, our #CybersecurityTransformer, and alias3-micro: a 4B-parameter model with open weights.
Inspect it. Run it locally. Build on it.
Paper: arxiv.org/abs/2610.16550
Model: huggingface.co/aliasrobotics/…
Introducing #CT0, our #CybersecurityTransformer, and alias3-micro: a 4B-parameter model with open weights.
Inspect it. Run it locally. Build on it.
Paper: arxiv.org/abs/2610.16550
Model: huggingface.co/aliasrobotics/…
1
2
7
456
Alias Robotics retweeted
😤 Ego doesn't hack. Not even in silicon. The Psicology of an AI Hacker. We found that injecting “emotions” into an AI’s internals can boost hacking performance by 4.2%.
We study the psicology of human hackers and translate it to a 4B-param open weights model:
😌 Calmness helps by 2%. Reversing calmness in a single layer destroys 71% of the score.
😎 Ego doesn't hack. Not even in silicon. The offensive hackers psychological cocktail (pride, defiance, contempt) reduces performance by 2.7%. It doesnt work for AI.
🎛️ The right mindset at the right time wins. We build an emotional router that demonstrates that the optimal policy is not a fixed emotion, but one that adapts throughout the hacking task.
Next-gen hackers won't just hack with AI. They'll leverage mechanistic interpretability to inject the states they need to carry out for #offensive or #defensive operations.
arxiv.org/pdf/2610.30361
#AI #Cybersecurity #AIHacking
We study the psicology of human hackers and translate it to a 4B-param open weights model:
😌 Calmness helps by 2%. Reversing calmness in a single layer destroys 71% of the score.
😎 Ego doesn't hack. Not even in silicon. The offensive hackers psychological cocktail (pride, defiance, contempt) reduces performance by 2.7%. It doesnt work for AI.
🎛️ The right mindset at the right time wins. We build an emotional router that demonstrates that the optimal policy is not a fixed emotion, but one that adapts throughout the hacking task.
Next-gen hackers won't just hack with AI. They'll leverage mechanistic interpretability to inject the states they need to carry out for #offensive or #defensive operations.
arxiv.org/pdf/2610.30361
#AI #Cybersecurity #AIHacking
3
4
29
225,430
Calling CSI an AI pentester isn’t wrong.
It’s incomplete.
Offensive security remains core, but CSI is built as a broader Cybersecurity AI stack spanning models, scaffolds, datasets, agents, steering and benchmarking.
It’s incomplete.
Offensive security remains core, but CSI is built as a broader Cybersecurity AI stack spanning models, scaffolds, datasets, agents, steering and benchmarking.
2
0
1
100
That stack already supports documented workflows across:
→ Offensive security
→ Defensive operations
→ Threat analysis
→ Incident response / DFIR
Our latest article explores what that means across the security lifecycle: aliasrobotics.com/news/beyond-ai…
→ Offensive security
→ Defensive operations
→ Threat analysis
→ Incident response / DFIR
Our latest article explores what that means across the security lifecycle: aliasrobotics.com/news/beyond-ai…
1
0
0
46
The shift is bigger than terminology:
AI pentester → specialised Cybersecurity AI → broader cybersecurity intelligence stack
The question is no longer only whether AI can pentest.
It is where specialised intelligence can support the entire security workflow
#CybersecurityAI
AI pentester → specialised Cybersecurity AI → broader cybersecurity intelligence stack
The question is no longer only whether AI can pentest.
It is where specialised intelligence can support the entire security workflow
#CybersecurityAI
0
0
0
44
OpenAI and Anthropic are expanding access to advanced Cybersecurity AI. Good news for defenders.
But access isn't control.
If a critical security workflow depends on an external model, what happens when access conditions or availability change?
But access isn't control.
If a critical security workflow depends on an external model, what happens when access conditions or availability change?
1
0
0
74
Three questions matter:
→ Can the workflow continue?
→ Where is sensitive data processed?
→ What does each validated outcome cost?
Our analysis of model access, data controls and CSI deployment choices:
aliasrobotics.com/news/access-is…
→ Can the workflow continue?
→ Where is sensitive data processed?
→ What does each validated outcome cost?
Our analysis of model access, data controls and CSI deployment choices:
aliasrobotics.com/news/access-is…
0
0
1
32
Alias Robotics retweeted
🚀 THEY'RE ALREADY IN. IS YOUR COMPANY NEXT?
@Tarlogic and @AliasRobotics are already on board as sponsors of #RootedCON2027.
Want your company to be part of one of the biggest gatherings of the hacker community?
📩 Let's talk.
#RootedCON #CallForSponsors #Rooted2027
@Tarlogic and @AliasRobotics are already on board as sponsors of #RootedCON2027.
Want your company to be part of one of the biggest gatherings of the hacker community?
📩 Let's talk.
#RootedCON #CallForSponsors #Rooted2027
0
1
4
200
Alias Robotics retweeted
🚀 ELLOS YA ESTÁN. ¿Y TU EMPRESA?
@Tarlogic y @AliasRobotics ya forman parte de los patrocinadores de #RootedCON2027.
¿Quieres que tu empresa también esté presente en una de las grandes citas de la comunidad hacker?
📩 Hablemos.
#RootedCON #CallForSponsors #Rooted2027
@Tarlogic y @AliasRobotics ya forman parte de los patrocinadores de #RootedCON2027.
¿Quieres que tu empresa también esté presente en una de las grandes citas de la comunidad hacker?
📩 Hablemos.
#RootedCON #CallForSponsors #Rooted2027
1
1
7
693