---
id: "contrarian-overrides-not-malicious"
type: "contrarian-insight"
source_timestamps: ["§ The Takeaways for Managers", "¶16"]
tags: ["cybersecurity", "prompt-injection", "user-behavior"]
related: ["claim-overrides-signal-design-flaws", "action-reframe-overrides"]
challenges: "The cybersecurity view that prompt injection and guardrail bypass attempts are inherently malicious or adversarial."
sources: ["tail1"]
sourceVaultSlug: "hbr-seg-tail1"
originDay: 1
articleStem: "hbr-tail-113-ai-personality-problem"
sourceUrl: "https://hbr.org/2026/06/does-your-ai-have-a-personality-problem"
sourceTitle: "Does Your AI Have a Personality Problem?"
---
# Prompt Injection as a Usability Symptom

**Challenges:** The cybersecurity view that prompt injection and guardrail-bypass attempts are inherently malicious or adversarial.

**The reframe:** In cybersecurity and AI-safety communities, prompt injection and attempts to bypass system guardrails are almost universally viewed as adversarial attacks. This research reframes these actions *in the workplace context*: when ordinary employees try to trick the AI, it is often a **desperate usability workaround** to neutralize a hostile or unhelpful system — not sabotage. This is the interpretive layer over [[claim-overrides-signal-design-flaws]], operationalized in [[action-reframe-overrides]] and captured by [[quote-ai-fighting-them]].

*Balance caveat (from enrichment):* a mature stance treats frequent overrides as **both** a usability signal *and* a security concern — triggering design review *and* risk assessment, not purely benign reframing.