A Google today announced the suspension of the creation of images of people in its generative Artificial Intelligence (AI) instrument Gemini, after “recent problems” with this functionality.
Since then, criticism has emerged on social media pointing out errors about some images created, in particular regarding gender and diversity in historical events.
Several users highlighted that Gemini was underrepresenting white people in the generated images.
"We are looking to resolve recent issues with Gemini's imaging functionality. In the meantime, we will suspend the generation of images of people and, soon, we will make an improved version available“, Jack Krawczyk, the product director at Google who is running the Gemini project, told AFP.
Since the end of 2022 and with the success of Chat GPT, generative AI, capable of producing all types of content [text, sound, image or video] at a simple request formulated in common language, has sparked a massive mobilization and all major operators in the technological sector have become involved in the development of instruments to organizations and individuals.
Increasingly developed, this technology has also led to similar developments in the alert system.
Last week, twenty of the most advanced companies in this field, including Meta, Microsoft, Google and OpenAI, announced their commitment to developing new techniques to identify content with lies or bias produced using AI.
