ChatGPT-5, Gemini 2.5, and Claude 4.5 have offered different predictions about which jobs will disappear due to automation. A study by Northwestern University and American University reveals that this lack of consensus calls into question the reliability of AI exposure indices that politicians and business leaders use to make key decisions.
The technical disagreement behind the algorithms 🤖
Economists observed that each model prioritizes different variables when calculating automation risk. ChatGPT-5 focuses on repetitive tasks and data analysis, Gemini 2.5 weighs social interaction, and Claude 4.5 highlights creativity and complex problem-solving. This technical divergence creates biases that prevent a unified index, making its application in public policies or business strategies difficult.
Spoiler: not even the AIs know what will become of your job 🎭
It turns out that asking an AI to predict your job future is like asking three drunk oracles: each gives you a different answer, and none get it right. While economists argue, office workers can breathe easy: if even the models themselves cannot agree, perhaps the only thing that will soon disappear is blind faith in automatic forecasts.