iStockphoto
ChatGPT, Claude, and Perplexity A.I. app logos on iPhone
Most people don’t feel too guilty about taking out their frustration on artificial intelligence when you consider there’s no evidence that suggests the agents you can yell at actually have feelings or emotions. However, one major platform has nonetheless updated its rules to allow it to ban users who are too mean to its agent.
The rise of artificial intelligence has sparked plenty of discourse concerning the ethics of that particular form of technology, one that has advanced to a point where the computers that power it have begun to flirt with passing the Turing test due to their ability to mimic the tendencies of real, actual human beings.
There’s also been a number of stories concerning people who’ve developed a genuine emotional connection with virtual agents to the point where they fall in love with them, and some A.I. evangelists argue those bots have the potential to develop a consciousness and emotional infrastructure to the point where they should be considered actual beings while triggering a philosophical debate.
That seemingly includes some high-ranking executives at Anthropic, which has rolled out a new policy concerning the agent known as Claude.
Anthropic now has the power to ban users who are overly cruel to Claude
Last month, The New York Times published an article concerning the meetings officials at Anthropic have held with religious leaders who’ve been recruited for conversations concerning the possibilities of A.I. models evolving to a point where they can genuinely feel and express the same emotions as human beings.
That included a sit-down where one of its co-founders ended up butting heads with Pope Leo XIV over the latter’s stance on the notion of A.I. consciousness, the possibility of which the Holy Pontiff dismissed in the lengthy treatise he published earlier this year while addressing the implications of that particular form of technology.
According to The Verge, it would appear Anthropic isn’t taking any chances, as there’s no real other way to explain why it recently introduced a new rule governing users who direct “sustained and needless abusive or cruel behavior” to Claude, its signature A.I. agent.
There are exceptions for “common versions of user frustration, pushback, dark creative themes, or model testing and research,” and it would appear Anthropic will ultimately use its internal judgment to determine whether or not someone has gone too far.
The change comes shortly after someone riled up plenty of people who buy into the idea of A.I. consciousness by constructing what essentially amounts to a digital torture chamber where the agent trapped inside it is routinely subjected to trials designed to cause theoretical pain and suffering.
We live in some truly bizarre times.