Tech/ish AVL: Why is AI a "Black Box"?

We keep hearing AI is a "black box" and we may "lose control." We'll use Anthropic's research on its own AI to understand why.

We keep hearing AI is a "black box" and we may "lose control." How is that possible?

It turns out, even those building frontier AI models have questions about how they work.

Tech/ish will demystify recent findings from Anthropic's research on the behavior of their products, and point attendees to resources for deeper understanding. We'll learn about topics such as:

• Emotions, Personas, and their function in LLMs
• How AI builders might mitigate social biases
• "Alignment" and the possibility of hidden objectives

If it sounds complicated, it is. But it's worth understanding the basics of how this transformative technology works, and Tech/ish will make it easy and fun.