Google DeepMind has released SL2T, an AI model that translates sign language directly to text in real-time, now available on Pixel 11 phones via Gboard and Live Transcribe. Trained on 100,000+ hours of sign language data across 50+ languages, SL2T uses computer vision to track full-body movements and employs machine translation rather than word-by-word conversion—addressing the 70 million deaf and hard-of-hearing users globally. The breakthrough ditches earlier flawed glossing approaches and on-device processing keeps raw video private.