6 min read
|
Saved February 14, 2026
|
Copied!
Do you care about this?
The article argues against the notion that judgment is a uniquely human skill. It highlights that AI can already make complex judgments and learn from feedback, challenging misconceptions about AI's capabilities. The author explores why these misconceptions persist, emphasizing that both AI and humans have strengths and weaknesses in judgment.
If you do, here's more
Yann LeCun, Meta’s Chief AI Scientist, made headlines in early 2022 when he claimed that AI models like GPT-3 couldn’t grasp real-world physics. He thought such models would fail to predict outcomes involving physical interactions, like a plate on a table when the table is pushed. Just two months later, GPT-3.5 demonstrated otherwise. This misjudgment highlights a broader issue: many experts misjudge AI's capabilities, particularly regarding complex skills like judgment, which an Op-Ed in the New York Times recently labeled as uniquely human.
The Op-Ed author argues that judgment involves navigating competing values and making decisions in ambiguous scenarios, claiming AI lacks this ability. However, the article counters this by showing that AI can already perform complex judgments effectively. For instance, AI models can analyze how interpersonal dynamics, like CEO hostility, might impact acquisition prices. Studies published in reputable journals such as Science and Nature found that AI outperformed humans in mediation and ethical dilemma scenarios, indicating that AI's judgment capabilities are already comparable, if not superior, to those of humans.
A significant reason for the misconception about AI’s limits is that critics often rely on outdated models, which can lead to skewed perceptions of AI's true capabilities. Many don’t consider the context or specific models in their evaluations. The Op-Ed’s author used a model that likely wasn’t state-of-the-art. Moreover, the AI's inability to factor in CEO relationships stemmed from a lack of proper contextual information rather than a deficiency in judgment. Humans still have advantages in recognizing subtle social cues, but AI developers are working to bridge these gaps.
Finally, there’s a tendency to romanticize human abilities and resist the idea that AI could match or exceed them. Many cling to the belief that human judgment is inherently special, perhaps out of a desire to feel unique in the age of AI. However, evidence from surveys suggests that AI-generated art can be as compelling as that created by humans, challenging the notion that human creativity holds a unique spark. The ongoing evolution of AI suggests that its capacities, including judgment, will continue to improve, warranting a reassessment of our expectations.
Questions about this article
No questions yet.