PHELMS™

Safety, control and traceability for the interaction between people and high-risk artificial intelligence.

 

It is the difference between an artificial intelligence that acts and one that can prove why it acted.

what is PHELMS™

PHELMS is a safety layer that sits between whoever decides (a person, an artificial intelligence system or both) and the machine that acts: robots, drones, vehicles or remotely operated machinery. 

 

It does not replace the manufacturer or the operator: it is added to what already exists so that no action is executed without authorisation, so that the person keeps command and so that everything can be proven afterwards.

«Authorizes: no action without verified permission»

 

«Keeps the person in control: the machine proposes; the person decides»

 

«Stops first: the stop command always prevails»

 

«Proves: every decision is recorded so that it cannot be altered without being noticed»

why now

Europe requires proof of effective human oversight of artificial intelligence systems and autonomous machines. Anyone operating robots, drones or AI in real operations will need evidence, not promises.

Milestones: 

 

«9 December 2026 · Directive (EU) 2024/2853 on liability for defective products»

 

«20 January 2027 · Regulation (EU) 2023/1230 on machinery»

 

«December 2027 · high-risk obligations under Regulation (EU) 2024/1689, the AI Act».

for whom

«Private security and surveillance with robots and drones»

 

«Service and industrial robotics»

 

«Autonomous vehicles and platforms»

 

«Remote operations in critical environments»

train and evaluate

Technology alone is insufficient if the supervisor is unprepared. PHELMS includes a training and evaluation method for human supervisors of automated systems, with realistic scenarios and documented results.

PROJECT STATUS

«Private security and surveillance with robots and drones» · «Service and industrial robotics» · «Autonomous vehicles and platforms» · «Remote operations in critical environments»

CONTACT

Controller: Fermín Sánchez Bernal. Purpose: to reply to your enquiry. Rights: access, rectification, erasure and others, at info@phelms-ai.com. Full details on the Legal page. By ticking this box I confirm I have read it and consent to my data being stored and processed for that purpose; I can withdraw my consent at any time.*

* Indica los campos obligatorios
¡Gracias! Nos pondremos en contacto con usted lo antes posible.

PHELMS es un proyecto de I+D con solicitudes de patente y de marca en trámite ante la OEPM. Queda prohibida la reproducción de su documentación sin autorización escrita. © 2026 PHELMS

www.phelms-ai.com · info@phelms-ai.com

Information icon

Necesitamos su consentimiento para cargar las traducciones

Utilizamos un servicio de terceros para traducir el contenido del sitio web que puede recopilar datos sobre su actividad. Por favor revise los detalles en la política de privacidad y acepte el servicio para ver las traducciones.