Jason Ross

2 talks in the archive.

Filters+

Materials

Has slidesHas repo

Difficulty

Track

Topics

Common in these results

Length

Year

1–2 of 2 talks

Sort

Talks·50 min

Jason Ross

Lays out a repeatable method for manipulating LLM agents by borrowing social engineering tactics like authority and pretext.

aisocialmachine-learning+2

Talks·51 min

Jason Ross

Surveys generative AI risks for enterprises, from prompt injection and data poisoning to model theft, with mitigation ideas.

aimachine-learningrisk+2