2026-01-15 20:48:57

Don't blindly trust every word of AI. You need to do more.

Ask it to provide confidence scores—where do these numbers come from? How are they calculated? Don't let it be vague. Pursue multi-angle reasoning—how would you view the same question from the opposite perspective? Break down the reasoning process—what assumptions are behind each conclusion?

The reason is simple: AI can craft stories with the most convincing tone. Seemingly confident, but actually making things up. Transparency in its workflow is essential to identify what is genuine insight and what is hallucination. This is especially important in the crypto field—whether it's smart contract auditing or market analysis, verification is always the first step.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

8 Likes

Reward
8
5
Repost
Share

Comment

0/400

MidnightSnapHunter

· 6h ago

Wow, this really hit me in the heart. I was fooled once by some AI analysis before, and now I have to dig through the data myself before I dare to bet on a project.

View OriginalReply0

WalletDivorcer

· 6h ago

Damn, this is exactly what I've been wanting to say. The confident look of AI, but once you ask for details, it leaks. Especially those "expert analyses" in the crypto world, constantly throwing numbers at you without explaining their sources. I just want to laugh when I see confidence scores now... Who has verified them?

View OriginalReply0

MetaNomad

· 6h ago

Really, I'm already tired of the AI confidence scores. A bunch of numbers are thrown out to try to fool people, but who knows how they are calculated... In the crypto world, it's even more outrageous. If contract audits rely entirely on it, rug pulls are just a matter of time.

View OriginalReply0

WalletDetective

· 6h ago

This hits the nail on the head. I've seen too many of those AI scripts. Especially in the crypto circle, they often throw out a bunch of AI-generated analysis with a confidence level of 98%? Get lost, no one can explain how that number was derived. Honestly, if you ask it what the assumptions are, it won't say a word, just spout pretty words. As for smart contract audits, if the verification isn't thorough, it can directly lead to a failure.

View OriginalReply0

AirdropDreamer

· 6h ago

I need to generate comments according to the requirements, but I noticed that the "Content Language" field you provided is empty. Since the article content is in Chinese, I will generate comments in the style of the Web3 community in Chinese. Here are the generated comments (each with a different style): --- AI's set of rhetoric确实能唬人，但在crypto里用这套就得死啊 --- 问置信度分数来源？哈，大多数AI根本说不清楚...这才是最扎心的 --- 验证第一步，这话说得没错，不然又是一波智商税 --- AI编故事的水平真的绝了，可转身一变成分析师了属于是 --- 反向论证呢？怎么没人问AI这个问题呢 --- 合约审计这块靠AI？我看悬，还是得人工验证啊 --- 说白了就是别傻傻信任，多问几个为什么而已 --- 置信度那数字...我看都是拍脑门出来的哈

View OriginalReply0