July 10, 2024  3:25 PM GMT-4 – 4:05 PM GMT-4
Truth & AI

How Are AI Companies Dealing With the Problem of Truth?

If you put an infinite number of monkeys in a room, they might write Shakespeare. Turns out that if you put a near-infinite number of parameters on a server, an LLM can too. In both cases, the answers might look right—but are they? How can we decide if algorithms produce true representations of the world when we can't even agree on it ourselves?
Legitimacy is the currency of the digital world. It's a demonstration of earned trust, without which others will not attend to, or act upon, our ideas. But online, generative content, anonymous sources, and alternate facts risk making every idea illegitimate.
The Internet is making us question the very nature of truth, from philosophical notions like objectivism and science, to Orwellian fears of revisionist history where tyrants retcon their sins. In this panel, we'll tackle the nature of truth in the modern world: What it is, what tools we have to defend it, and what we must build to ensure a functioning public square for future generations.

Founder and CEO
Waverly
Senior Machine Learning Engineer
Ground News
Founder
The Society Library
Principal Manager, Startup Programs
Mila – Quebec Artificial Intelligence Institute
Notice at collection Your Privacy Choices