Should the legal rights be obtained?

In one paper ELEOS AI was published, non-profit non-profit equipment for assessment and consciousness using “computer functionalism”. A similar idea used to be or sprayed any other of the paths, although criticized later in career. The The theory proposes These human minds can be considered specific types of computer systems. From there you can understand if other computer systems, like Chabot, have indicators of feelings similar to those of man.

Eleos AI said in the newspapers that “a great challenge in implementing” this approach “to involve significant calls to the judgment, both in formulating indicators and assess their presence or absence in AI and systems.”

The welfare model is, of course, initial and still evolving field. There are a lot of critics, including Mustaf Suleiman, General Manager of Microsoft AI, which is recently Posted a blog About “Apparently aware AI”.

“This is too early and sincerely dangerous,” Suleiman wrote, which generally referred to the field of welfare research model in the model. “All this will exacerbate, create even more dependency problems, prey on our psychological vulnerabilities, introduce new dimensions of polarization, complicate the existing rights for rights and create a huge new category of error for society.”

Suleiman wrote that “today there is zero evidence” that there is aware of AI. He turned on the connection to a paper This long is 2023 long, which suggested a new framework for assessing whether and the system has “properties of indicators” of consciousness. (Suleiman did not respond to a comment request from the wired.)

I talked Campbell for a long time soon after Suleiman published my blog. They told me that they agreed more than what he said, they do not believe that they do not believe that the model of welfare research should stop existing. Instead, they claim that damage Suleiman are the correct reasons are referenced why They want to study the theme first.

“When you have a big, confusing problem or question, one way to guarantee that you will not solve her to throw your hands and be like” Oh Wow, this is too complicated “. “I think we should at least try.”

Consciousness testing

The model of exploring researchers primarily concerns the issues of consciousness. If we can prove that you and I and I are aware that the same logic could be applied to large language models. In order to be clear, or long or Campbell thinks it is aware today, and they are not sure that they will ever be. But they want to develop tests that would allow us to prove it.

“Miscellances are from people who deal with the actual question,” Is this, conscious? “And that he has a scientific framework for thinking about it, I think it’s just robustly good,” Long says.

But in a world where and research can package in sensational titles and videos of social media, in the main philosophical issues and the experiments for bending the mind are easily wrongfully deserved. Take what happened when anthropic has published a Security Report This showed Claude Opus 4 can alleviate “harmful actions” in extreme circumstances, such as blackmail invented engineer to prevent it from turning it off.