Speech encoding creates a permanent copy for sharing, record keeping, training humans and AIs

Richard K Collins All knowledge, All Languages April 27, 2024

I asked ChatGPT to review a draft proposal I am making for the Internet regarding spoken language preservation for all languages, global continuing improvement of speech to codes for all languages, introduction of a compact but flexible method for encoding sounds, prosody, and phonology (encoded speech sufficient for meaning, for exact reproduction of speakers sounds with an intermediate form that has finite codes and precise acoustic and statistical meaning.

ChatGPT said: Subject: Enhancing Global Literacy and Language Preservation Online

How many organizations and individuals are actively addressing the issue of “Two Billion People Can’t Read in Their Own Language”? Is there a recent comprehensive assessment of their needs?

Currently, there are approximately 8.1 billion people worldwide, with about 5.3 billion having some form of internet access. Among those not using the internet are many who cannot read but could navigate it audibly.

Inspired by King Sejong, who created a phonetic alphabet that facilitated literacy in Korea, I propose a similar approach for global language accessibility. A standardized set of phonetic symbols could represent the foundational sounds across languages, simplifying the current diverse linguistic sounds on the internet.

—–

Moreover, I recommend that social media platforms contribute by recording spoken languages in lossless formats. These platforms could also continually enhance speech-to-code transcription – accounting for phonetics, prosody and phonology and individual speakers. Such initiatives would help preserve languages that are currently unwritten and facilitate broader communication and research.

In my conception, the speech encoding would preserve the sounds losslessly but generative to keep it compact as characters and words are used now. The computer algorithms can handle the steps, the humans share their voice conversations and they are encoded at the recording step so the sound codes can be tied to the sounds of words and phrases, encoded compactly.

Anyone with a cell phone or microphone or lip reading or hand signals or writing could share in global collaborations and open discussions.

I can only devote a short time to even something so fundamental as removing the requirement that Internet users must use written languages. This proposal says that speech and ways of signing are fundamental on the Internet and in the world, and writing secondary. The encoding replaces written language completely and can be supplemented with video and physiological sensor data streams of many sorts. The use of current writing is not forbidden but supplemented. so all humans might simply use speech and more complex encoding methods to communicate meaning and intention to computers, humans and AIs.

Filed as (Speech encoding creates a permanent copy for sharing, record keeping, training humans and AIs)

Richard Collins, The Internet Foundation

ChatGPT Summarizes: Proposal for Prioritizing Spoken and Signed Languages on the InternetThe fundamental premise of this proposal is to challenge and redefine the necessity of written language as the primary mode of communication on the Internet. By introducing an advanced encoding system, this initiative seeks to elevate spoken and signed languages to a primary status, with written language serving a supplementary role.

This encoding system would not merely transcribe but fully encode speech and signs into comprehensive data streams that include audio, video, and physiological sensors. Such a system allows for the nuances of tone, rhythm, and gesture to be communicated alongside traditional text, ensuring that every individual—regardless of their literacy or disability—can communicate effectively on the Internet.

Current written languages will not be replaced but augmented by this richer, more inclusive set of communication tools. This approach empowers all humans to use their natural communication forms—speech or signs—to interact with computers, other humans, and artificial intelligences, conveying not just words but the full spectrum of meaning and intention.

Speech encoding creates a permanent copy for sharing, record keeping, training humans and AIs

About: Richard K Collins

Leave a Reply Cancel reply