As part of strengthening the enhancement of “decentralized infrastructure” utilizing smartphones towards the “social implementation” of 5G/web3, a technical preview has been announced for the collaboration between the “freebit Edge LLM,” an AI system operating on smartphones, and the family monitoring service “TONE Family” provided by Tone Mobile. This preview aims to allow AI to assess the risk level of children’s social media usage.
Freebit has been dedicated to promoting decentralization by establishing “decentralized infrastructure” utilizing smartphones, striving to realize the true form of the internet. In 2009, Freebit introduced “ServersMan,” a web server service operating on smartphones, which achieved the top position in the App Store’s category ranking. In 2022, Freebit released “TONE Chain (TONE Coin),” an Ethereum-compatible L1 blockchain operating on smartphones, reaching the third-largest node scale worldwide. Building upon the evolution of technology, Freebit has now developed “freebit Edge LLM,” an AI system operating on smartphones, as the third installment following web servers (Edge www) and blockchain nodes (Edge blockchain), and has made it available as a technical preview through Tone Mobile’s service.
Unlike traditional filtering systems that assess SNS notification texts based on keyword databases, “freebit Edge LLM” detects the connections themselves using AI operating on smartphones and assesses the risk level. It plans to notify the guardians using the mechanism of “TONE Family.” During this risk assessment, no information is sent to external servers, and the risk level is determined solely within the device, aiming to achieve extremely high privacy. Furthermore, for those seeking even higher precision analysis, “freebit Edge LLM” can collaborate with a larger-scale generative AI system within Freebit’s data center to conduct high-precision analysis (in this case, data is sent externally with user consent).
Freebit intends to further enhance the accuracy of “freebit Edge LLM” by implementing mechanisms such as user-participatory learning data enhancement through the web3-type stakeholder community verification experiment “One Vision.
