 What is the history of AI and speech processing? The start for speech processing was back in the 1950s and 60s where single digits and words were recognized. Around the 70s, that is after Hidden Markov models were introduced, the complex problem of speech to text could be broken down into simpler problems, language modeling, pronunciation modeling, and acoustic modeling. But using these three models creates its own complexities. By the 2000s, with the availability of rich data sets, neural approaches started taking over. They could directly learn to convert speech to text and they were hence more accurate. And today, we make use of models like GPT where a pre-trained language model is fine-tuned on speech tasks. So we need less speech-to-text related data. Subscribe for more on artificial intelligence.