10/30/2026
The old advice was to watch for bad spelling in a phishing email. What happens when the attacker has an AI writing assistant too?
NIST's Generative AI Profile, AI 600-1, addresses it in §2.9, Information Security. It names two primary risks. G*I could lower the barriers for, or ease automated exercise of, offensive cyber capabilities. And it expands the attack surface, because G*I itself is vulnerable to attacks like prompt injection or data poisoning.
On the first, NIST says offensive capabilities advanced by G*I may augment attacks such as hacking, malware, and phishing.
The profile is voluntary and requires nothing of your practice. Its one specific link to Texas law is narrow: under Tex. Bus. & Com. Code §552.105(e)(2)(D), a defendant may not be found liable for a violation it discovers through an internal review process while substantially complying with the most recent version of the profile or another recognised AI risk framework.
What is required already sits elsewhere. For a business completing income tax returns, the FTC Safeguards Rule, 16 CFR §314.4(e)(1), requires security awareness training for your personnel, updated as necessary to reflect risks identified by your risk assessment.
Want the plain-English map of what a Texas tax practice owes? Type briggssocial.com/tax into your browser, or email [email protected] and we will send it over.
https://briggssocial.com/tax