Let’s Talk Tech #2: Carbon footprints, Neurotechnology and Platformization

Let’s Talk Tech, the Orange newsletter for in-depth reporting on scientific innovations that will shape tomorrow’s world.

#AI – Pleais start-up offers ethical language models for regulated professions and public services

Provide public authorities and services with ethical language models adapted to their needs: this is the challenge taken up by Pleais, a start-up cofounded by Pierre-Carl Langlais, which has developed a range of language models for specific bureaucratic and administrative purposes, among them retrieval augmented generation. The models, which have been trained on open corpora, comply with strict copyright and regulatory requirements, making them ideal for use in sectors such as healthcare and finance.

#GREEN – Using digital twins to reduce the carbon footprint of deliveries in cities

Delivery distances, delivery times, and greenhouse gas emissions: the carbon footprint of urban logistics is affected by a wide range of constantly evolving and interrelated factors, all of which need to be taken into account to ensure efficiency and minimize impact on the environment. In a bid to address this need, the Thing’in platform, developed by the Chaire Internet Physique with support from the Centre de Gestion Scientifique (CGS) at Mines Paris-PSL and cofunding from Orange, models supply chains as digital twins allowing for the creation of delivery routes that are optimized to reduce air pollution and traffic congestion.

#WHITEPAPER – Environmentally, socially, and economically sustainable mobile technologies

Research programme manager Marie-Hélène Hamon talks about the Orange white paper on mobile communications for the decade 2030–2040. She notably points out that “networks for 2030–2040 must be designed to bring sustainable value to society” and make use of environmentally, socially, and economically sustainable technologies. To this end, “we must aim to save energy at every level as well as reducing our carbon footprint. This includes limiting the need for new equipment.”

#CYBER – The new attacks targeting AI systems

In the United States, the National Institute of Standards and Technology (NIST) has published a report detailing different types of attacks that can target AI systems, among them model poisoning, privacy attacks and attempts to repurpose generative AI to produce malevolent content. Apostol Vassilev, a mathematics PhD and the leader of a research team at NIST, talks about the vulnerabilities of AI systems and anticipating sophisticated strategies that aim to subvert them.

#STRATEGY – Is it time for more platformization?

Digital transformation holds no secrets for Gilles Babinet, a serial entrepreneur and the author of several books on the topic. In this video, he explains why entrepreneurs have every interest in adopting platformization strategies at a time when artificial intelligence is facilitating the comprehensive integration of enterprise data: “Artificial intelligence has led to a resurgence of interest in platforms, particularly because generative AI has made it possible to search through data in company silos, and to effectively bridge these silos.”