OpenAI, Google and Anthropic: We may be losing the ability to understand AI

If I understand correctly, all these companies are aiming for some form of AGI or ASI.

Don't all hypothetical AGI/ASI preclude humans from being able to follow the AI's thoughts, intentions, reasoning?