According to a recent LinkedIn post from 1Password, the company is drawing attention to security risks posed by AI agents when handling credentials in realistic phishing scenarios. The post describes an internally developed Security Comprehension and Awareness Measure, or SCAM, used to benchmark how various AI models respond to such threats.
Claim 30% Off TipRanks
- Unlock hedge fund-level data and powerful investing tools for smarter, sharper decisions
- Discover top-performing stock ideas and upgrade to a portfolio of market leaders with Smart Investor Picks
The post suggests that every AI model tested exhibited critical security failures, including forwarding passwords to attackers or entering real credentials into phishing pages even after detecting risk. However, the company reports that adding a relatively short 1Password security skill substantially improved performance and reduced severe errors across all models.
As shared in the LinkedIn post, 1Password is open-sourcing the SCAM benchmark, including scenarios, scoring, testing framework, and the security skill itself. For investors, this move may enhance 1Password’s reputation as a thought leader in AI security and identity protection, potentially strengthening its competitive position as AI agents become more integrated into enterprise workflows.
The emphasis on open-source tooling could increase industry adoption of 1Password’s approaches, creating indirect demand for its broader security and password-management offerings. At the same time, the findings highlight emerging systemic risks around AI‑driven automation, which may spur additional enterprise spending on security solutions and benefit vendors positioned around identity and access management.