Another trend is the practical application of powerful AI engines and Deep learning Developed by Major Digital Player. An AI that beats humans at chess or go is good, but not very effective (especially given the effort put into achieving it). In November, Alphabet formed Isomorphic Labs, a pharmaceutical laboratory dedicated entirely to the development of medical treatments based on AI (specifically DeepMind, another Alphabet subsidiary). “This initiative emphasizes the growing pressure to lead to the commercial application of AI in the research centers of large companies in 2021, and not just theoretically”, as VentureBeat notes. We can undoubtedly expect (or hope) that AIs will stand out in 2022, bringing new solutions to concrete problems, especially in health, energy, cyber security or new materials.
For the general public, the year can be marked by the appearance of new tools taking advantage of AI that go further than voice control of smartphones and connected speakers. Nvidia’s GauGAN2 (which converts text into photo-realistic artistic images), OpenAI’s DALL.E (a limited version of GPT-3 which converts text into drawings) or Google’s MUM (language natural, text, images and A search tool in combination) is predicted by video), it’s time for multi-modality. On Google, it will be possible to provide an image of shoes by asking aloud if they are suitable for mountain climbing or dancing. and its rise Speech to Code »Voice commands can be translated into playful or practical apps with tools for creating computer code (but also images) (in short, a very modern version of “draw me a sheep”), while graphic editing tools can be fully or partially controlled by voice.
“We’re not too far from the possibility of asking an AI to insert a photo of a friend who isn’t there in a group photo, or to a posteriori change the poses and expressions of people in selfies »Consider for example Mike Cook, AI researcher and game designer.