Openai launches a program for “field” standards for “special” for “field”

Openai, like many artificial intelligence laboratories, believes that artificial intelligence standards are broken. She says she wants to fix it through a new program.

The program is called the Openai Pioneers program, it will focus on creating assessments of artificial intelligence models Blog post.

“With the acceleration of the pace of adopting artificial intelligence through industries, there is a need to understand and improve its impact in the world,” the company continued its publication. “The creation of the field Evals is one of the ways to better reflect real use, which helps the difference to evaluate the typical performance in practical environments with high risk.”

as recently Controversy With the standard standard lm Arena and Meta’s MAVERICK model, it is difficult to know, these days, exactly what distinguishes one model. It measures many artificial intelligence criteria used widely on internal tasks, such as resolving mathematics problems at the level of PhD. Others can control, or do not match well with most people’s preferences.

Through the Pioneers program, Openai hopes to create standards for specific areas such as legal, financing, insurance, health care and accounting. The laboratory says that in the coming months, it will work with “multiple companies” to design designed standards and exchange these standards in the end publicly, as well as “industry” assessments.

“The first group will focus on startups that will help in laying the foundations of Openai Pioneers”, “Openai wrote at the blog post. “We choose a handful of startups for this initial regiment, each of which works on high -value -valuable cases where AI can pay the effect of the real world.”

In the program, companies in the program will also get the opportunity to work with the Openai team to create typical improvements by controlling Fine Tuning, a technology that improves models for a narrow set of tasks, says Openai.

The big question is whether the artificial intelligence community will embrace criteria for funded by Openai. Openai has previously supported financial measurement efforts, and designed its own assessments. But partnership with clients may be seen to issue artificial intelligence tests as a far -fake bridge.