Despite its impressive output, generative AI doesn’t have a coherent understanding of the world

[A] popular type of generative AI model can provide turn-by-turn driving directions in New York City with near-perfect accuracy — without having formed an accurate internal map of the city.

Despite the model’s uncanny ability to navigate effectively, when the researchers closed some streets and added detours, its performance plummeted.

When they dug deeper, the researchers found that the New York maps the model implicitly generated had many nonexistent streets curving between the grid and connecting far away intersections.

This could have serious implications for generative AI models deployed in the real world, since a model that seems to be performing well in one context might break down if the task or environment slightly changes. —news.mit.edu

Post was last modified on %s = human-readable time difference 3:14 pm

Share
Published by
Dennis G. Jerz
Tags: ai

Recent Posts

Will Journalism Be a Crime in a Second Trump Administration?

Journalism is not a crime. It's not a crime if it offends the powerful. It's…

6 days ago

I just caught myself thinking, “This doesn’t suck.” #medievalyork #mysteryplay #blender3d #design #aesthetics

I just caught myself thinking, “This doesn’t suck.” #medievalyork #mysteryplay #blender3d #design #aesthetics

1 week ago

A spooooooky post about predatory journals for this Halloween season.

Predatory publications are not concerned with writing quality (or even coherence), and thus also do…

1 week ago