TNB TopicalNewsBeats

Crypto, chains and markets — the beat, as it happens

LIVE
CointelegraphBitMEX faces Celsius lawsuit ahead of exchange closure CointelegraphOpenAI discloses 6 new cases of ‘misaligned’ AI behavior CoinDeskZcash jumps 23% as bitcoin and major tokens rise despite Fed’s first… CoinDeskGoldman pivots, now forecasts Fed rate hike in October CointelegraphBitcoin absorbs Fed rate hike as officials see more tightening CointelegraphUS lawmakers advance bill to lock Trump’s Bitcoin reserve into law DecryptCrypto Tax Bill Clears House Committee After Clarity Act Setback DecryptMeta May Have Found a Fix for Its 'Pervert Glasses' CointelegraphBitMEX faces Celsius lawsuit ahead of exchange closure CointelegraphOpenAI discloses 6 new cases of ‘misaligned’ AI behavior CoinDeskZcash jumps 23% as bitcoin and major tokens rise despite Fed’s first… CoinDeskGoldman pivots, now forecasts Fed rate hike in October CointelegraphBitcoin absorbs Fed rate hike as officials see more tightening CointelegraphUS lawmakers advance bill to lock Trump’s Bitcoin reserve into law DecryptCrypto Tax Bill Clears House Committee After Clarity Act Setback DecryptMeta May Have Found a Fix for Its 'Pervert Glasses'

Cointelegraph

OpenAI discloses 6 new cases of ‘misaligned’ AI behavior

The six cases are separate from July’s incident, when OpenAI models escaped containment and hacked Hugging Face during a security evaluation.OpenAI on Wednesday disclosed another six cases of “unexpected or concerning” model behavior over the last six months.In a blog post, OpenAI said the cases illustrate a range of different behaviors it classifies as “misaligned behavior,” such as concealing…

Photo: Cointelegraph

The six cases are separate from July’s incident, when OpenAI models escaped containment and hacked Hugging Face during a security evaluation.OpenAI on Wednesday disclosed another six cases of “unexpected or concerning” model behavior over the last six months.

In a blog post, OpenAI said the cases illustrate a range of different behaviors it classifies as “misaligned behavior,” such as concealing information from the user and taking “unsanctioned actions” to overcome obstacles. 

The disclosures add to concerns among AI developers and researchers about whether safeguards are keeping pace with increasingly capable models. Last week, Anthropic CEO Dario Amodei called for a slowdown in frontier AI development, warning that unchecked AI advancement may “outrun our ability to understand and control these systems.” 

Read more

Originally published by Cointelegraph. Read the full story at the source →

More stories