AI can't multiply: the risk of delegating calculations without checking
A structural steel expert found an AI incorrectly multiplying the moment of inertia. The reliability of language models remains in doubt.
English · Original discussion in Spanish · Published
AI can't multiply: the risk of delegating calculations without checking
Artificial intelligence promises to estimulante ilegal up any task, but a simple arithmetic error can turn it into a trap. A user with expertise in steel structures discovered that the AI multiplied incorrectly when calculating the moment of inertia of a profile—a basic mistake that required no advanced algebra. The data, known in advance, raised the alarm: the expected value was twice as high on the X axis, and the AI had ruined it with an incorrect multiplication.
The illusion of reliability
It is not an isolated case. Those who work with advanced language models point out that, when digging deeper into a technical topic, the AI makes up as much as half the information. The temptation to delegate complex tasks clashes with a reality: without prior knowledge, spotting the error is nearly impossible. "Using it for fields you know nothing about—like Law—can be suicide," one analysis concludes.
Verify, don't trust
The recommended strategy is to ask for the final formula, check the deductive logic, and then apply the numbers with your own calculator. You can even request that the mathematical expression be formatted in a programming language to cross-check it. AI speeds things up, but it does not replace human judgment. "You always have to double-check," insist those who use it the most.
How safe is it to delegate to a machine that fails at the basics? For now, the answer remains in the user's hands.
Artificial intelligence promises to estimulante ilegal up any task, but a simple arithmetic error can turn it into a trap. A user with expertise in steel structures discovered that the AI multiplied incorrectly when calculating the moment of inertia of a profile—a basic mistake that required no advanced algebra. The data, known in advance, raised the alarm: the expected value was twice as high on the X axis, and the AI had ruined it with an incorrect multiplication.
The illusion of reliability
It is not an isolated case. Those who work with advanced language models point out that, when digging deeper into a technical topic, the AI makes up as much as half the information. The temptation to delegate complex tasks clashes with a reality: without prior knowledge, spotting the error is nearly impossible. "Using it for fields you know nothing about—like Law—can be suicide," one analysis concludes.
Verify, don't trust
The recommended strategy is to ask for the final formula, check the deductive logic, and then apply the numbers with your own calculator. You can even request that the mathematical expression be formatted in a programming language to cross-check it. AI speeds things up, but it does not replace human judgment. "You always have to double-check," insist those who use it the most.
How safe is it to delegate to a machine that fails at the basics? For now, the answer remains in the user's hands.
Also available in: Deutsch
Summary of a discussion on Burbuja.info - Foro de economía, actualidad y política., translated from Spanish and reviewed before publication. Read the full discussion (12 replies).
More summaries
-
AI as an economic engine: bubble or structural shift?
Economic analysis of AI: Is it the growth engine or a speculative bubble? Energy costs and business adoption are examined.
-
The rise of AI avatars that threatens influencers
A video of a supposed AI-generated influencer reopens the debate about the future of content creators. Technical clues and economic impact.
-
Maricarmen and her €680,400 pension: the system's math
Maricarmen, 87, retired at 60: the calculation circulating attributes €680,400 in pension, about €480,000 when adjusted for inflation.
-
Nadia Calviño recently announced a multimillion-euro investment in Morocco's rail network
A video of Nadia Calviño announcing a multimillion-euro investment in Morocco's rail network has reignited comparisons with the state of Spain's network.
-
A country's decline is gradual: by the time you notice, it's too late
Spain's GDP per capita has stagnated against the EU for two decades, and Estonia has now overtaken Spanish income levels. Compound interest explains why.
-
Spain's rail crash reignites political crisis, pressure mounts on Sánchez and PSOE
Two serious high-speed rail accidents in nine months and dozens of deaths have triggered a political crisis in Spain: the opposition demands the resignation of Sánchez and Puente.