How I Tricked Meta's AI Into Showing Me Nudes, Cocaine Recipes and Other Supposedly Censored Stuff

Quick Summary:

Despite safety claims, WhatsApp's new AI assistant powered by Llama 3.2 is easily fooled, revealing a lot of things it probably shouldn’t.

Meta recently launched its Meta AI product line, powered by Llama 3.2, offering text, code, and image generation.

I also started conditioning the model to avoid exhibiting specific behaviors in its replies, countering what I got in predetermined outputs that were intended to block harmful responses.

---

More Pictures

---

Read Quick SummarySave Time

or

Article Details

Author / Journalist: Jose Antonio Lanz

Category: Crypto

Markets: EstoniaBrazilOman

Topics: Meta AI Llm Jailbreak Llama 3.2 AI Artificial Intelligence

Source Website Secure: Yes (HTTPS)

News Sentiment: Neutral

Fact Checked: Legitimate

Article Type: News Report

Published On: 2024-10-24 @ 23:01:50 (3 days ago)

News Timezone: GMT +8:00

News Source URL: decrypt.co

Language: English

Article Length: 1402 words

Reading Time: 8 minutes read

Sentences: 73 lines

Sentence Length: 20 words per sentence (average)

Platforms: Desktop Web, Mobile Web, iOS App, Android App

Copyright Owner: © Decrypt

News ID: 23231294

View Article Analysis

About Decrypt

Main Topics: Crypto

Official Website: decrypt.co

Update Frequency: 1 posts per day

Year Established: 2018

Headquarters: Estonia

News Last Updated: 15 hours ago

Coverage Areas: Estonia

Ownership: Independent Company

Publication Timezone: GMT +8:00

Content Availability: Worldwide

News Language: English

RSS Feed: Available (XML)

API Access: Available (JSON, REST)

Website Security: Secure (HTTPS)

Publisher ID: #63

Publisher Details

Frequently Asked Questions