It would take really smart and highly educated people to use these tools to make bad things (I'm participating in authoring an article about how to develop screening systems to prevent this.) It doesn't seem likely that one or even a few loners who want to do harm will even think of taking this route. (And why would they?) As you note, guns and explosives are cheap and plentiful.
It is hard to stay calm when AI chaos gripped everyone/all around. Very well 👌🏻 written and informative piece. Lots of food for 💭 thought. Reminded us about Milton Leitenberg again.🤔
This article feels like a straw man argument. It certainly doesn't feel like a fair representation of concerns raised and recommendations made by people in the space, which actually tend to be along the lines of "pandemics can be devastating, AI *could* (not will) increase those risks, we should learn more about this potential threat".
In addition, as you yourself admit, you don't have direct experience using the latest AI models (e.g. your viewpoint comes from "...reviewing many stories as to the inadequacy of AI-generated products").
And the "basic hypothesis" that you posit describes capabilities of many LLMs currently available (its now a running joke in the AI community that if a sceptic says "An LLM will never be able to do X" its almost certainly the case that an LLM already can).
You also misrepresent the arguments of Hersman and Nelson, because they're asking some of the same sensible questions that you are, and many people in the community ask the very same sensible questions you do as well.
It would take really smart and highly educated people to use these tools to make bad things (I'm participating in authoring an article about how to develop screening systems to prevent this.) It doesn't seem likely that one or even a few loners who want to do harm will even think of taking this route. (And why would they?) As you note, guns and explosives are cheap and plentiful.
It is hard to stay calm when AI chaos gripped everyone/all around. Very well 👌🏻 written and informative piece. Lots of food for 💭 thought. Reminded us about Milton Leitenberg again.🤔
This article feels like a straw man argument. It certainly doesn't feel like a fair representation of concerns raised and recommendations made by people in the space, which actually tend to be along the lines of "pandemics can be devastating, AI *could* (not will) increase those risks, we should learn more about this potential threat".
In addition, as you yourself admit, you don't have direct experience using the latest AI models (e.g. your viewpoint comes from "...reviewing many stories as to the inadequacy of AI-generated products").
And the "basic hypothesis" that you posit describes capabilities of many LLMs currently available (its now a running joke in the AI community that if a sceptic says "An LLM will never be able to do X" its almost certainly the case that an LLM already can).
You also misrepresent the arguments of Hersman and Nelson, because they're asking some of the same sensible questions that you are, and many people in the community ask the very same sensible questions you do as well.