>the company actually is ethical and safety conscious everywhere

Anthropic is emphatically not safe. None of the AI labs with customers (i.e., excluding a few small nonprofits whose revenue comes from donations) are anything like safe -- because of extinction risk. The famous positive regard that Anthropic employees have for their organization's mission means almost nothing because there have been hundreds of quite destructive cults and political parties whose members believed that theirs is the most ethical and benign organization ever.

The best thing you can say about Anthropic is that if you have to support some AI lab by becoming a customer, investor or employee, it is slightly less dangerous for the world to support Anthropic than OpenAI although IMHO (and I admit I am in a minority on this among extinction-risk activists) it is slightly less dangerous to support Google Deep Mind or Mistral than Anthropic.

All four organizations I mentioned should be shut down tomorrow with their assets returned to shareholders.

The current crop of services provided by the leading AI labs are IMHO positive on net in their effect of people and society, but the leading AI labs are spending a large fraction of the 100s of billions of dollars they've received from investors on creating more powerful models, and they might succeed in their goal of creating models that are much more powerful than the ones they have now, which is when most of the danger would manifest.

The leaders of all of the leading AI labs have the ambition of completely transforming society and the world through AI.