artificial intelligence

A team has developed a wearable designed to aid navigation for people who are blind or visually impaired. The system uses artificial intelligence (AI) algorithms to probe the environment and send signals to the wearer when approaching an obstacle or object. The technology, which is presented in Nature Machine Intelligence, was tested with humanoid robots and blind and partially sighted participants in both virtual and real environments.

According to a new study published in Science, a machine learning-based artificial intelligence (AI) system - called Mal-ID - can decipher an individual's history of infections and diseases in the immune system. The authors say this provides a powerful tool with the potential to accurately diagnose autoimmune disorders, viral infections and vaccine responses.

An artificial intelligence (AI) model led by the company Meta is capable of translating speech and text, including direct speech-to-speech translations, from up to 101 languages in some cases. According to the research team, this model - called SEAMLESSM4T - can pave the way for fast universal translations ‘with resources to be made publicly available for non-commercial use’. The work is published in the journal Nature.

The development of generative artificial intelligence, and in particular of large language models, could generate between 1.2 and 5 million tonnes of accumulated electronic waste between 2020 and 2030, according to a study published in Nature Computational Science. The study estimates the mass of waste generated by hardware elements such as processing units, storage units and power supply systems.

A study published in the journal Nature describes a tool capable of inserting watermarks into text generated by large linguistic models - artificial intelligence (AI) systems - thereby improving their ability to identify and track artificially created content. The tool uses a sampling algorithm to subtly bias the model's choice of words, inserting a signature that can be recognised by the detection software.

Two of the Nobel Prize winners in Chemistry 2024 are employees of Google DeepMind, who caused significant unrest among their colleagues in May. Hassabis and Jumper announced in *Nature* the results of their AlphaFold 3 model, with applications in drug design; however, they published it in a closed manner, with reviewers not even having access to the system, which contradicts the basic principles of scientific publication. We risk having the transformative potential of AI controlled by big tech companies.

The Royal Swedish Academy of Sciences has awarded the Nobel Prize in Chemistry 2024 on the one hand to David Baker for computational protein design, which makes it possible to construct proteins with functions not present in nature. On the other hand, jointly to Demis Hassabis and John M. Jumper of Google DeepMind, for the development of AlphaFold2, which allows the structure of the 200 million known proteins to be predicted at high speed. 

The Royal Swedish Academy of Sciences has awarded the Nobel Prize in Physics 2024 to researchers John J. Hopfield and Geoffrey E. Hinton for discovering the foundations that enable machine learning with artificial neural networks. Hinton for discovering the foundational basis that enables machine learning with artificial neural networks. This technology, inspired by the structure of the brain, is behind what we now call ‘artificial intelligence’. 

Large language models - Artificial Intelligence (AI) systems based on deep learning, such as the generative AI that is ChatGPT - are not as reliable as users expect. This is one of the conclusions of international research published in Nature involving researchers from the Polytechnic University of Valencia. According to the authors, in comparison with the first models and taking into account certain aspects, reliability has worsened in the most recent models, such as GPT-4 with respect to GPT-3.

A study published today in Nature Genetics examines AstraZeneca's new tool, MILTON, which uses artificial intelligence to detect biomarkers and predict diseases before they are diagnosed. According to this analysis, the tool could potentially predict over a thousand diseases and may even be more effective than the currently available polygenic risk scores.