I haven't found a single peer-reviewed paper covering AI-powered productivity increase that also takes into consideration the additional work required to review, correct, and oversee AI outputs.
The "increased productivity" and "socially beneficial" claims have been at the core of major AI investments and government partnerships, and have been raised to justify legal exceptionalism (especially in fields such as copyright and data protection).
Yet, it's 2025, and there doesn't seem to be a single paper capable of objectively demonstrating what AI companies have been marketing over the past two and a half years (including the review/oversight time, which they conveniently ignore).
This is a sad state of affairs for AI governance.