If you are a platform provider dealing with disinformation and polarization — this project developed a value-aware toolbox that implements guardrails to stop asocial and immoral behavior.
Ethical AI Guardrails for Safer Social Media, Robotics, and Medical Decisions
Imagine giving an AI a moral compass instead of just a list of rules. Instead of just guessing what we want, this system can actually think about whether an action is right or wrong based on human values. It's like teaching a computer to have a conscience so it doesn't make dangerous or biased mistakes.
What needed solving
Current AI lacks a moral compass, leading users to attribute intent to machines that don't have it. This creates dangerous gaps in social media safety, robot ethics, and medical decision-making.
What was built
The Reflective Global Neuronal Workspace (RGNW) model and a corresponding prototype toolbox for embedding value-awareness into AI.
Who needs this
Who can put this to work
If you are a robot manufacturer dealing with potential manipulation or deceit in human-robot interaction — this project developed a model that constrains robot behavior within ethical norms.
If you are a software developer dealing with complex clinical choices — this project developed a decision-making support system that ensures medical protocols are value-aligned.
Quick answers
What is the cost or pricing for this technology?
Based on available project data, no pricing or cost structures have been disclosed as the project is in the development phase.
Can this be deployed at an industrial scale?
The project is currently prototyping a toolbox and testing it in 3 specific domains, but full industrial scale data is not yet provided.
What are the IP and licensing terms?
Based on available project data, specific licensing terms are not mentioned, though it is an EU-funded Horizon-EIC project.
How does this help with AI regulations?
It does not propose new regulations but embeds ethics directly into the AI technology to make it more reliable and trustworthy.
How is the system integrated into existing AI?
Integration is achieved through a toolbox and the Reflective Global Neuronal Workspace (RGNW) model that acts as a moral consciousness component.
Who built it
The consortium is well-balanced for technology transfer, featuring 7 partners across 5 countries. With an industry ratio of 43% (3 industrial partners, including 1 SME), there is a strong bridge between the 4 academic/research entities and real-world commercial application.
Contact AGENCIA ESTATAL CONSEJO SUPERIOR DE INVESTIGACIONES CIENTIFICAS
Talk to the team behind this work.
Contact us to explore licensing the RGNW toolbox for your AI products.