Searching...
18 results for “ai security”
Topic4 mentions across episodes
AI security
Concept3 mentions across episodes
security
Safety are making AI go well, and the risk management is just much more of a broader problem. It's got some technical aspects, but I I I I think that's a small part of it. I don't know that the leaders of the labs would say, like, we can do nothing a
3:02 / 4:13
from transcript3:02 – 4:13
...is agentic security. So if we can't even trust chatbots to be secure, how can we trust agents to go and book us flights, manage our finances, pay contractors, walk around embodied in humanoid robots on the streets? You know, if somebody goes up to a
56:10 / 57:19
from transcript56:10 – 57:19
Speaking about, AI is assisting with, like, command and control. I mean, I I I remember was hearing some story about, how on Wall Street, humans used to not be able to you always had a human in the loop for each decision. So at a later stage before t
12:05 / 13:07
from transcript12:05 – 13:07
but we can all work together. Amazing. And I love that. Is there anything that you think will never be placed in the realm of proactive AIs? I think if you take all the risks that we have talked about, in AI and AI safety, that just becomes amplified
24:45 / 25:52
from transcript24:45 – 25:52
...if you're a security organization in a big company, you look at this and go, oh my god. How do I protect my employees, my customers, my data, my models, my developers, my IT? We look at it as all the same problem. How do I adopt AI in a safe, confide
2:14 / 3:15
from transcript2:14 – 3:15
So the the best thing to do is to use the scientific method to do more research to try and, more precisely define those risks and, of course, address them. And I think that's what we're doing. I think there probably needs to be, 10 times more effort
2:06:40 / 2:07:48
from transcript2:06:40 – 2:07:48
I'm almost surprised how much effort all the bigger AI labs are putting into trying to to ret team a lot of this and, like, trying to make sure that AI basically is somewhat aware of when a goal that it's given leads to things that it shouldn't be do
42:37 / 43:50
from transcript42:37 – 43:50
using new technology, in this case, general purpose technology and repurposing it for harmful ends. And that's a huge, risk. And I think that has a lot of complications because, generally, you know, I'm in huge favor of open science and open source.
2:07:29 / 2:08:32
from transcript2:07:29 – 2:08:32
safety research should never be done in your opinion for AI? I have a list for biology that I'm you know, like, I don't think you should pass certain viruses through mammalian cells to make them more infectable or do gain a function mutations on them
32:56 / 34:12
from transcript32:56 – 34:12
to discrimination, like hiring and firing and whether insurance claims are approved or not. If you look look at all those risks with AI, they get amplified, in the agentic world. Right? And if we don't do it right, and the agents can be very, very fa
25:24 / 26:30
from transcript25:24 – 26:30
by and large, you know, the way cyber vendors kind of promote their products is around fear, uncertainty, and doubt, which is if you don't do something, something bad's gonna happen. And for the first time, we could actually use the word enablement,
5:48 / 7:02
from transcript5:48 – 7:02
assuming they don't, like, invent a new bioweapon or something like that. But with agents, there's all types of bad stuff that can happen. And if you deploy improperly secured, improperly data permission to agents, people can trick those things into
18:44 / 19:48
from transcript18:44 – 19:48
multiple forms of safety in AI. And I think people kind of mix the terms to mean different things. Right? One form of it is, is it AI somehow being offensive or crude or, you know, using language you don't like or concepts you don't like? There's a s
25:25 / 26:39
from transcript25:25 – 26:39
by other humans. If you date another human being, if you start, you know, being violent to that human being, there's gonna be consequences. But the issue with AI and future robotics and all that is the the morality around it. You you can do whatever
1:08:53 / 1:10:01
from transcript1:08:53 – 1:10:01
...cybersecurity people can do a great job with that. And then there's there's a third a third option here, which is maybe you need a a system that is both truly agentic, and can also be tricked into doing bad things by a malicious user. There are some
1:00:00 / 1:01:00
from transcript1:00:00 – 1:01:00
You can go and and kinda train it against that, but you can never be certain with any strong degree of accuracy that it won't happen again. This does start to feel like a little bit like the Aliven problem where, like, in theory, you know, it's like
1:16:26 / 1:17:30
from transcript1:16:26 – 1:17:30
Have a podcast?
Run your podcast through Clypt
Get ranked clips, hooks, and ready-to-post copy from your own episodes. Free to try.