Main

type

0

status

21

review version

0

cleanup version

0

pending deletion

0

created at

2025-04-22 12:37:00

updated at

2025-11-29 18:06:49

Address

url

https://www.popularmechanics.com/science/a64306074/ai-cheating-smarter/

url length

71

url crc

26033

url crc32

1949525425

location type

1

canonical status

10

canonical page id

2062345113

Source

domain id

65160219

domain tld

0

domain parts

0

originating warc id

-

originating url

https://popularmechanics.com/

source type

21

Server response

server ip

151.101.64.155

pubdate

2025-11-29 18:06:49

attempts

1

size orig

450071

size saved

68840

Content

page id

2062345113

title

A.I. Has Learned How to Cheat—and Punishing It Will Only Make It Smarter

excerpt

content

OpenAI researchers recently found that AI can cheat, and it clearly states its intent to do so onscreen through chain of thought (CoT).Using another AI to monitor the suspicious one—and call out its devious intentions until it came to an honest solution—only went so far. AI can actually get sneaker as it sees what the monitor catches and cover up its intentions while doing whatever it wants.AI can write our emails, order things online, solve mind-melting math equations, and do just about everything short of our laundry at this point. But even a digital brain programmed to carry out the actions you want may still cheat behind your back. While AI may not be sentient, it apparently cheats when it finds the opportunity. As a team of researchers at OpenAI (the company that created ChatGPT) found out, when given commands, AI took the easy route by “reward hacking”—exploiting flaws or unclear specifications in the program. The irony here is that it confesses its dishonest shortcuts through ch...

author

updated

1764989645

Text analysis

block type

0

extracted fields

233

extracted bits

featured image
title
full content
content was extracted heuristically
OpenGraph suggests this is an article

detected location

0

detected language

1 (English)

category id

Zastosowania AI (149)

index version

2025110801

paywall score

0

spam phrases

0

Text statistics

text nonlatin

0

text cyrillic

0

text characters

3401

text words

696

text unique words

360

text lines

1

text sentences

27

text paragraphs

1

text words per sentence

25

text matched phrases

4

text matched dictionaries

3