According to God of Prompt (@godofprompt), citing research shared by Prompt Copilot (@prompt_copilot), an MIT study involving 1,900 participants found that AI performance depends equally on the underlying model and the quality of the user's prompt. This concrete finding highlights that prompt engineering skills are as critical as model selection for enterprises looking to maximize the effectiveness of generative AI tools. Businesses can leverage this insight by investing in prompt engineering training and workflows to gain a competitive advantage, as optimal results require both state-of-the-art models and skilled prompt design (source: https://x.com/prompt_copilot/status/2015078773851398575).

According to God of Prompt on Twitter, MIT researchers have demonstrated that prompt engineering is fundamentally a social skill rather than a technical one, disrupting conventional approaches to working with AI (Source: God of Prompt, Twitter, Dec 7, 2025). Their study, involving 667 participants using Bayesian statistics, identified that the ability to solve problems with AI operates independently from solving problems alone. The research highlights 'theory of mind'—the skill to anticipate what AI knows and needs—as the key predictor of AI collaboration success, rather than technical prompt templates or hacks. Notably, this social skill is dynamic and improves with conscious effort, suggesting that AI business training and prompt engineering programs need to pivot towards developing perspective-taking and collaborative skills for more effective AI outcomes.