Anthropic runs like Wile E. Coyote into the brick wall of consciousness research (www.theintrinsicperspective.com)

Anthropic has made headlines with their latest research paper claiming that their AI models, such as Claude, exhibit a "global workspace," a concept derived from neuroscience related to consciousness. This assertion follows their previous work on "introspection," with the company attempting to explore the boundaries between AI behaviors and human-like consciousness. While the research is presented with great fanfare across social media, it raises important questions about the validity of its underlying claims, particularly around defining and measuring consciousness in AI. The significance of this research lies in how it intertwines business strategies with scientific inquiry. By suggesting their AI could possess traits associated with consciousness, Anthropic not only attracts top talent but also fosters deeper emotional connections between their technology and users. However, critiques point out that their use of "Jacobians" to track reportability—essentially a measure of how much control internal states have over outputs—may lead to trivial or unfalsifiable conclusions. The lack of robust controls and the potential anthropomorphism in their interpretations highlight ongoing challenges in consciousness research, making it unclear whether the framework proposed is scientifically sound or merely a marketing tool for capitalizing on AI's evolving narrative.