Tether CEO Paolo Ardoino suggests localized AI models as a solution to prevent hacks in blockchain. Ardoino argues that locally executable AI models are capable of protecting privacy and independence.
Modern devices have sufficient power for this approach, according to Ardoino. He stated, "Smartphones and laptops can fine tune general large language models (LLMs) with user's own data, preserving enhancements locally to the device."
Ardoino described this as a "paradigm shift" in user privacy and independence.
"By running directly on the user's device, be it a smartphone or laptop, these models eliminate the need for third-party servers. This not only ensures that data stays local, enhancing security and privacy, but also allows for offline use," Ardoino said.
Tether's boss also emphasized that users could maintain control over their information.
According to him, Tether, issuer of the leading stablecoin USDT, is exploring the integration of locally executable models into its AI solutions.
The proposal follows a hack at OpenAI. Reports indicate a hacker accessed OpenAI's internal messaging systems in early 2023 and compromised details of the company's AI design. That was before ChatGPT was launched to a massive success.
And that was not the only security scandal around security issues with the most successful AI company. Recently, ChatGPT conversations of macOS users were found to be stored in unencrypted plain-text files. That was a huge deal, as many users share very sensitive information on those chats, including their own private financial data and work info.
Apple recently announced plans to integrate ChatGPT with its "Apple Intelligence" AI-powered products. The encryption issue has reportedly been resolved. However, questions remain about why it occurred.