We are somewhere on the hype curve of AI and companies such as Google, Microsoft, Anthropic, and OpenAI are pushing it hard because of financial incentives and regulatory capture. Everything we’re being told right now is not the reality of these models.
Are models like Claude, ChatGPT, et al a genuine threat? Yes, I do believe they are but concerns from governments and organizations are misplaced.
In capable hands it can be a genuine threat because it’s still a very powerful tool, however incapable hands can also be dangerous with it but for the opposite reasons. It presents new vectors for vulnerability because the public is grossly misled by the capabilities of these products and this overuse it to a point where they add vulnerabilities and bugs into whatever is made by them.
"AI" as we know it do not have the autonomy that humans or other animals have. A human can be good natured but suddenly snap and go rogue; a human can change careers and (tasks) jobs on a whim. AI cannot do that without a significant level of human interaction.
Lastly and a major reason imho: adoption of the products is grossly low because the tech industry as a whole have been disconnected from the reality public perception. They’re desperate to hype it up even if it means targeting malicious actors to use it.