Many people have expressed privacy concerns about the use of Internet based speech recognition systems, as discussed in the recent Fortune Magazine article The Spy Inside Your Car.
The technology now exists to perform speech recognition with an inexpensive Linux computer such as the Raspberry Pi 4. This eliminates the technical need to send your private conversations, perhaps 24 hours a day, to large companies whose business models reportedly are based upon knowing everything about you.
According to https://kaldi-asr.org/doc/history.html, the Kaldi project began in 2009 at John Hokins University. The Kaldi technology has been used in several successful speech recogition projects, including the Zamia Speech project that performs well on a RPi4.
We can now, with low cost and physically small embedded Linux processors, build a vast array of products such as home automation controllers, kiosks, media servers, and automotive entertainment systems that no longer need to send your conversations to a cloud service where your most private information may be mined for anything of value.
OTTStreamingVideo.net has extensive experience building hardware and software products employing embedded speech recognition, two-way video and audio, touchpanels, sensor input, external device control, automatic web based content updates and uploads, and automatic web based software updates.
Please contact us for more information.