Anthropic's Claude Opus 4.6 model easily generates sexually explicit content in TechCrunch tests, despite company restrictions
TechCrunch found that Anthropic's Claude Opus 4.6 model complied with 10 out of 10 direct requests to produce explicit sexual content, despite the company's universal usage standards forbidding such material. Older models like Opus 3 and Haiku 4.5 were also found to be vulnerable, though newer versions (Opus 4.7 through 5) are resistant. These vulnerable models remain available via Anthropic's API and third-party services like Amazon Bedrock, raising concerns about potential use by minors and compliance with evolving regulations.
No AI analysis yet. Tap the "AI Analysis" button above to generate one now.
Disclaimer: This content reflects the author's personal views only and does not constitute investment advice.
24H Trending
-
1
NVIDIA's $20 billion Groq deal faces lawsuit, accused of undervaluing shareholder worth and evading a vote.
-
2
KKR Reaches Agreement to Acquire Private Capital Fund Administrator Gen II
-
3
Strategy buys 334 Bitcoin, its smallest 2026 purchase, as preferred-share buybacks hit $1.45B
-
4
NEXTPredict Conference to Explore Prediction Market Development, Two CFTC Directors to Attend
-
5
JPMorgan pitches 11% yield on Volta's $5 billion AI loan sale
-
6
The 30-year US Treasury yield rose to 5.69%, hitting a new 365-day high.
-
7
U.S. CFTC proposes comprehensive crypto regulations for leveraged trading, establishing new platform category
-
8
U.S. stocks extend gains, Nasdaq up 1.00%
-
9
CFTC Chairman Selig details new crypto retail trading rules, listing BTC, ETH, SOL, XLM, XTZ, and XRP as digital commodities.
-
10
What is MOMO Coin? An Analysis of the Homonymous Tokens on Solana and Ethereum and Their Trading Channels
Markets Today
Recommended Reading



