Significant step forward in AI research

New Aurora-M model focusses on safety

2024/04/22 by

Computer science doctoral student Felix Friedrich is engaged in the development of a novel AI model that prioritises security, in addition to free accessibility and multilingualism. Friedrich is conducting research as part of the TU project “Reasonable Artificial Intelligence (RAI)”, which is under consideration for designation as a Cluster of Excellence in the Excellence Strategy of the German federal and state governments.

A collaboration of international researchers has developed the new AI model Aurora-M, which is not only freely accessible and multilingual, but also focuses on safety. As part of this global initiative, Felix Friedrich (PhD student in the Artificial Intelligence and Machine Learning research group under Professor Kristian Kersting at TU Darmstadt) has concentrated on “safety tuning” (optimising safety). In doing so, he ensures that the AI not only corresponds to our values, but is also protected against misuse. The US Biden-Harris Executive Order on Artificial Intelligence served as a valuable reference point for optimising safety. This legal text places a clear focus on the safety and accountability of AI technologies and serves as a reference to establish a democratically legitimised standard for our values and safety.

The AI model is based on open source.

Another distinctive feature of the novel AI model is its complete open-source availability and its capacity to process multiple languages. “This step is pivotal, as AI technologies are increasingly influencing our daily lives and it is imperative that they are accessible to a diverse range of individuals from diverse linguistic and cultural backgrounds,” states Friedrich. The The utilisation of previously marginalised languages, such as Hindi, Vietnamese or Finnish, in particular represents a significant improvement in accessibility. By disclosing the source code, the developers are enabling researchers from around the world to learn from and build upon the progress made, which will ultimately further advance the development of AI technology as a whole.

Part of the “RAI” Cluster of Excellence

There is an important connection to the “Reasonable Artificial Intelligence (RAI)” project, which has been invited to submit a full application in the “Cluster of Excellence” funding line as part of the Excellence Strategy of the German federal and state governments. RAI has set itself the goal of developing and promoting AI technologies that are not only effective and efficient, but also prioritise ethical principles such as safety, fairness and transparency. The new AI model is a direct result of these efforts and highlights the TU's dedication to responsible AI research and development.