Despite the media’s recent portrayal of AI agents as autonomous entities capable of human-like interactions, it is crucial to frame these AI systems as tools under human supervision.
Can’t emphasize this enough. Tools are tools. As much as their devs want them to be just tools for X use by Y people, they can be abused,misused and otherwise “hacked” to be more than those artificial restrictions.
Criticisms are also of implementation, not the technology or its capabilities as a whole.
AI will never be at the level as a human, and shouldn't be treated as such
In many specific applications current AI algorithms already exceed human capabilities. That's been the case for much longer than the LLM craze. And as new AI algorithms are developed it becomes more and more likely that even without Artificial General Intelligence, the category of computing known as AI could reach a point that it exceeds human capabilities in all possible applications.
But even with that fact it's true that we shouldn't treat the current implementations of AI as if it were human.
If we ever develop an actual AGI and it is ad capable as a human, then we should reconsider, but nothing we have now seems to imply we are close to that.