urn:li:share:7049501759660068866

Do you know anyone can create security/bias attack #bias #security in plain english against #chatgpt #generativeai #llms as opposed to using sophisticated algorithms? ChatGPT can be used for Personalized malicious content generation, bypassing its content filters #! The data you add into ChatGPT is aggregated by OpenAI and used for further building out its state of knowledge. An excerpt from my latest post.. Prompt Injection: Security attacks on LLMS that can be carried out in plain English. Gives new meaning to low code/no code programming.