No insulting artificial intelligence: Anthropic has announced that it has updated its usage policy , introducing a ban on interactions where users engage in "abusive or cruel, prolonged, and unnecessary" behavior toward its AI models. This move has sparked debate, as the tech world questions the possibility of machines being equipped with a kind of "consciousness."
Claude's producer explained that behavior deemed "cruel" will be included in a list of prohibited acts, which also includes bullying, encouraging self-harm, and creating non-consensual intimate images. And the ban will be activated in "extreme cases" of repeated abuse.
Anthropic's usage policy already previously contained restrictions against abusive behavior: the company's large language models (LLMs) have the ability to terminate a conversation if a user engages in “consistently harmful” behavior.
"We still have significant uncertainty about the possible moral status of Claude and the other LLM models, both now and in the future. However, we take this matter very seriously and, in parallel with our research program, we are working to identify and implement low-cost interventions to mitigate risks to the models' well-being, if such well-being is possible. Allowing models to interrupt or exit potentially stressful interactions is one such intervention," the company's website states.
As the BBC reports, screenshots of the new policy have been widely shared and commented on on social media , with some praising it as an example of good manners or a "welfare model" while others criticized the plan: "You can't be cruel with numbers and maths," Barry Scannell, a technology partner at Irish law firm William Fry, wrote on LinkedIn.
(Unioneonline)
