Tether’s Paolo Ardoino Warns Against Risks of Centralized AI Models

Tether CEO Paolo Ardoino recently expressed concerns on social media platform X about the risks associated with centralized large language models (LLMs).

Highlighting a major security breach at OpenAI in early 2023, Ardoino described the incident as “scary.” According to a report by The New York Times, OpenAI did not disclose the breach despite the exposure of sensitive information.

Former OpenAI researcher Leopold Aschenbrenner criticized the company’s security measures, suggesting they were inadequate against threats from foreign actors.

Aschenbrenner also alleged that he was dismissed from OpenAI due to political reasons, a claim the company denies, stating that the breach had been disclosed prior to his employment.

Concerns remain about OpenAI’s secrets potentially falling into the hands of foreign nations like China, despite assurances of no national security risks.

Beyond security issues, centralized AI models face criticism for unethical data usage and censorship. Ardoino advocates for the use of local AI models to address privacy concerns and ensure resilience and independence. He emphasized that modern smartphones and laptops are capable of fine-tuning general LLMs, making locally executable AI models a viable solution.