Who uses LLM prompt injection attacks IRL? Mostly unscrupulous job seekers, jokesters and trolls

Kaspersky examined nearly 1,000 web pages containing prompt-injection phrases (e.g., “ignore all previous instructions”) and found they are mainly used to game HR screeners, nudge product portrayals, express protest, or for jokes, rather than to enable destructive actions; notably, no malicious use in spam/scam campaigns was observed, and while research shows potential for phishing, container escapes, and data exfiltration, the firm assesses current real-world threat from prompt injection as limited and largely theoretical.