AI Standards Push Can Borrow From the Payments Playbook
Anthropic, Google and OpenAI may agree that powerful artificial intelligence models need common safety standards. Getting to that commonality will mean they must also set standards governing who writes the tests, who administers them, and what a model must do to fail. Payments knows that territory well, as shown by several years’ worth of initiatives. […] The post AI Standards Push Can Borrow…
Artificial intelligence developers from Anthropic, Google, and OpenAI have recognized the need for common safety standards for powerful AI models. The payments industry has already established a robust framework for such standards, which could serve as a valuable reference point. Payments organizations like the FIDO Alliance and EMVCo have developed specifications for passwordless authentication, contactless payments, and agentic payments.
These specifications define technical requirements, enable secure communication between different components, and establish processes for testing implementations. AI developers are now considering similar approaches to create standards for AI agents, including authentication, authorization, and evaluation of dangerous capabilities.
By taking a structured approach to establishing these standards, the AI industry can ensure the safe and secure deployment of increasingly powerful AI models.
Written by urgent.news from PYMNTS's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
