OpenAI stated that they might revise their safety measures if another firm introduces a high-risk AI model lacking protective features.

OpenAI wrote in its Preparedness Framework The report states that if another firm introduces a model posing a threat, the company might follow suit after "thoroughly" verifying that the "risk environment" has altered.

This document outlines the methods used by the company for tracking, assessing, predicting, and safeguarding against potential catastrophes caused by artificial intelligence models.

"If another front-line AI developer introduces a high-risk system without equivalent safety measures, we might revise our criteria," OpenAI stated in a blog post released on Tuesday.

Nevertheless, we would initially thoroughly verify that the risk environment has genuinely shifted, openly admit that we are implementing modifications, evaluate whether these changes do not substantially elevate the total risk of significant damage, and simultaneously maintain protections at a higher safeguarding threshold.

Prior to making a model available to the broader audience, OpenAI assesses potential hazards by pinpointing conceivable, quantifiable, novel, serious, and irreversible risks, and implementing protective measures for those threats. They categorize these risks into four levels: low, medium, high, or critical.

Certain risks that the company already monitors include their models’ abilities in areas such as biology, chemistry, cybersecurity, and self-enhancement.

The firm mentioned it’s also assessing fresh risks, including the capability of their AI system to operate continuously without human intervention, its potential for self-replication, and the threats this might present within the realms of nuclear technology and radiation.

"Risks associated with persuasion," like the use of ChatGPT for political campaigns or lobbying, will be managed separately from the framework and examined differently. Model Spec the document that defines ChatGPT's behavior.

‘Quietly reducing safety commitments’

Steven Adler, who previously worked at OpenAI, stated this during his interview X The revisions in the firm's readiness document indicate that it is "stealthily diminishing its pledges regarding safety."

In his posting, he highlighted the company’s pledge from December 2023 to test "finely tuned versions" of their artificial intelligence models. However, he also mentioned that OpenAI has decided to focus solely on testing models where the trained parameters or 'weights' will be made public.

"People might completely differ on whether fine-tuned model testing is necessary and preferable. It’s better for OpenAI to withdraw a pledge than to maintain it without adhering to it," he stated.

“However, in both scenarios, I would appreciate it if OpenAI could clearly communicate their departure from this earlier promise.”

The announcement follows the release of a new series of artificial intelligence models by OpenAI this week, known as GPT-4.1. system card Or safety report. Euronews Next reached out to OpenAI regarding the safety report but had not received a response by the time of publishing.

This follows the action of 12 ex-OpenAI staffers who have taken legal steps. brief Last week, in the lawsuit filed by Elon Musk against OpenAI, he claimed that transitioning into a profit-driven entity might result in compromises regarding safety measures.