Oxford Professor Flyvbjerg: ‘Current AI is more about persuasion than truth’

Bent Flyvbjerg is a Danish economic geographer and professor at the University of Oxford’s Saïd Business School and the IT University of Copenhagen. He is widely regarded as a leading global expert on megaproject management, risk, and decision-making, with his research influencing infrastructure, policy, and project planning around the world. Flyvbjerg is the author of several influential books, including Megaprojects and Risk and How Big Things Get Done, and has served as an adviser to government agencies, private corporations, and the World Bank.

Oxford professor emeritus Bent Flyvbjerg is warning that today’s artificial intelligence tools are being built to persuade rather than to tell the truth.

In his 2025 working paper, AI as Artificial Ignorance, Flyvbjerg writes, “AI and bullshit (in the strong philosophical sense of Harry Frankfurt) are similar in the sense that both prioritize rhetoric over truth. They mix true, false, and ambiguous statements in ways that make it difficult to distinguish which is which. AI sounds convincing even when it’s wrong.”

Flyvbjerg argues that large language models like ChatGPT are not dangerous because they are superintelligent but because they are super-persuasive while often inaccurate. Without a clear mechanism for identifying what is true and what is not, these systems can produce flawed or misleading outputs that appear credible.

He links this dynamic to Harry Frankfurt’s concept of “bullshit”—language that aims to impress or convince without regard for truth. According to Flyvbjerg, this is precisely how many AI tools currently function.

The result is what he calls “artificial ignorance,” where systems confidently generate information that lacks reliability. The paper includes several examples of AI hallucinations, including fabricated citations, inaccurate data, and persuasive but wrong conclusions.

Flyvbjerg urges researchers, technologists, and users to treat AI-generated information with deep skepticism unless it is independently verified. As AI becomes more embedded in education, journalism, and decision-making, he argues that critical thinking and truth-filtering mechanisms must be central to how these systems are deployed.

Otherwise, society risks confusing fluency with wisdom and rhetorical confidence with real knowledge.

Bent Flyvbjerg is a Danish economic geographer and professor emeritus at the University of Oxford’s Saïd Business School, and a professor the IT University of Copenhagen. He is widely regarded as a leading global expert on megaproject management, risk, and decision-making, with his research influencing infrastructure, policy, and project planning around the world. Flyvbjerg is the author of several influential books, including Megaprojects and Risk and How Big Things Get Done, and has served as an adviser to government agencies, private corporations, and the World Bank.

Leave a Reply

Your email address will not be published.