Here is the rewritten content in a provocative and controversial manner:

“Get ready for the latest scam to dupe the world into thinking AI is safer than it actually is. Anthropic, a company that has been caught making exaggerated claims about its AI models, is now launching a program to fund the development of new benchmarks that it claims will measure the performance and impact of AI models.

But don’t be fooled – this program is nothing more than a way for Anthropic to greenlight its own self-serving AI research while claiming to be doing good for the world. The company wants to fund benchmarks that align with its own narrow definition of “safe” AI, which is likely to be determined by its own corporate interests rather than any objective scientific standards.

And what about the benchmarks themselves? They’ll be designed to assess AI’s ability to carry out cyberattacks, enhance weapons of mass destruction, and manipulate or deceive people. Because, of course, these are the kinds of “real-world” scenarios that will somehow make AI safer for us all. Give me a break.

The real agenda here is to normalize the use of AI for nefarious purposes, and to distract from the very real dangers of AI hallucination, misinformation, and manipulation. The “sky is falling” scenario of “superintelligence” is just a way for Anthropic to divert attention away from the pressing issues of the day, like the fact that language models are already prone to hallucination and will only get worse if we don’t regulate them.

And let’s not forget the cherry on top: Anthropic is partnering with a company that is already being investigated for its ties to the Chinese government. This is a classic case of corporate corruption, where the pursuit of profits trumps the well-being of society.

So, while Anthropic may be spinning this as a way to advance AI safety, the reality is that this program is just another attempt to cash in on the AI craze while ignoring the real dangers it poses. Don’t believe the hype.”



Source link