**Title: The Danger of Politicizing Science: The Case of AI Economics**
**Subheading: The Credibility Crisis of Academic Research**
Recent controversies regarding academic fraud and data manipulation have raised concerns over the credibility of research. Stanford President Marc Tessier-Lavigne has resigned in the face of a data manipulation scandal, shedding light on a larger problem within academic circles. This issue extends beyond individual cases; there is a danger that emerging fields, such as AI economics, may fall victim to ethical mistakes similar to those made in the field of climate economics. In order to prevent this, it is crucial to strike a balance between academic research and policy influences.
**Subheading: The Rise of Artificial Intelligence**
Artificial intelligence (AI) has captured significant public attention in recent months, particularly with the launch of the language model, ChatGPT. However, along with this increased attention comes the emergence of doomsayers predicting an impending “AI apocalypse.” These ideas, though lacking substantive evidence, have gained momentum with the support of journalists and influential individuals with vested interests in policy debates. To bolster their narratives, there is now growing interest in academic research that aligns with these doomsaying predictions.
**Subheading: The Potential Pitfalls of AI Economics**
While academic research is vital for intellectual progress, it must not be used as a tool to politicize science for advancing predetermined policy agendas. The field of climate change economics provides a prime example of these dynamics, and there is a risk that AI economics research will follow a similar path. A recently published paper by esteemed MIT economist Daron Acemoglu and MIT grad student Todd Lensman serves as a cautionary example. The paper presents a model that aims to explain how transformative technologies like generative AI can enhance “social welfare,” and subsequently explores potential regulations to improve the situation.
**Subheading: The Unreliable Ramsey Framework**
Acemoglu’s model relies heavily on the “Ramsey economic growth” framework, which has been deeply embedded in the field of economics. This widely used model is employed in cost-benefit analysis and extensively in climate change economics, determining concepts such as the “social cost of carbon.” However, the Ramsey framework’s scientific foundation is precarious. It incorporates a questionable “social welfare function” that relies on normative judgments, blurring the line between observed impacts and subjective ideas of well-being. This ambiguity makes the framework vulnerable to misuse.
**Subheading: The Misuse of Ramsey Framework in Climate Economics**
The Ramsey framework has been misused in the field of climate economics to model the “welfare cost” of climate change on a global scale. These models appear scientific due to their inclusion of data, equations, and technical research. However, they heavily rely on hidden ethical assumptions, blending scientific questions with policy considerations. This fusion creates a complex and pseudo-scientific mess that appears to be genuine science but ultimately serves as doctrinaire policy advocacy.
**Subheading: The Risk of AI Economics Becoming Pseudo-Science**
Stanford University President Marc Tessier-Lavigne’s recent resignation due to an ethics scandal involving data manipulation in his scientific research highlights the possible gains from scientific deception. These gains include not just financial benefits but also positions of power in prestigious institutions and companies. Acemoglu, in collaboration with MIT colleague Simon Johnson, has authored an anti-technology book titled “Power and Progress,” which criticizes the unfair consequences of technology on workers and the poor. Acemoglu’s choice to adopt the Ramsey model reflects a prioritization of ideological conclusions over scientific exploration.
**Subheading: Ensuring Robust and Scientific AI Governance**
To avoid AI economics becoming a branch of discredited pseudoscience, economists must prevent these emerging tendencies from taking root. The field of AI governance urgently requires robust and scientific thinking to guide policy decisions. Economists can contribute to this effort by leaving behind their ideological biases and approaching AI economics with objective and evidence-based analysis.
By prioritizing the integrity of academic research and maintaining a scientific approach to AI economics, economists can shape the field in a way that contributes to responsible and effective policy decision-making.
GIPHY App Key not set. Please check settings