NEW YORK: Meta Platforms ( META.O ) said on Tuesday it will give researchers access to the components of a new “human-like” artificial intelligence model that it says is more accurate than existing models. Can analyze and complete incomplete images in a method.

The model, I-JEPA, uses background information about the world to fill in missing pieces of images, rather than looking at nearby pixels like other creative AI models, the company said.

The approach involves human-like reasoning advocated by Meta’s top AI scientist Ian Likon and helps the technology avoid errors that are common in AI-generated images, such as hands with extra fingers. .

Meta, which owns Facebook and Instagram, is a prolific publisher of open-source AI research through its in-house research lab. Chief Executive Mark Zuckerberg has said that sharing models developed by Meta researchers can help the company spur innovation, close security gaps and reduce costs. can be done

Read more: US Congress to Consider Two New Bills on Artificial Intelligence

“For us, it’s much better if the industry standardizes the core tools that we’re using and so we can benefit from other people’s improvements,” he told investors in April.

Company executives have dismissed warnings from others in the industry about the potential risks of the technology, with top executives from OpenAI, DeepMind, Microsoft ( MSFT.O ) and Google ( GOOGL.O ) last month. Refused to sign a statement in support of epidemics and its dangers with wars.

Lechon, considered one of the “godfathers of AI”, has spoken out against “AI doomerism” and argued in favor of building security checks into AI systems.

Meta is also starting to add creative AI features to its consumer products, such as advertising tools that can create image backgrounds and an Instagram product that can edit user photos, both based on text prompts.