iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
For U.S. federal employees, the safest baseline is to use AI for official work only through a tool their agency has approved for that purpose and the information involved. Treat prompts and uploads as disclosures, verify generated material before relying on it, and remain responsible for work or decisions made with AI assistance. Rules differ by agency, tool, data class, and task; state, local, contractor, and non-U.S. government employees must follow their own organization’s rules.
Start with your agency’s rules—not a general AI rule
There is no single employee policy that can safely be assumed to cover every U.S. federal agency, AI service, kind of information, and use. Check your agency’s current AI policy, approved-tool list, information-handling rules, and any requirements for the task before using a system for official work.
Approval must fit the specific use. An agency may distinguish public web services from enterprise or internally controlled tools, and may authorize a system for some information or purposes but not others. Do not assume that a tool available on the public internet—or one an agency uses somewhere—is cleared for your work.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsFor example, GSA’s directive, signed March 11, 2026 and updated June 12, 2026, sets AI assessment, procurement, usage, monitoring, and governance standards for GSA. Its applicability is specifically defined for GSA employees and contractors, relevant GSA systems and federal data, and AI systems acquired by or for GSA, subject to stated exclusions. It is an example of agency implementation, not a rule that automatically governs all federal employees. Read GSA’s directive.
#1 Best Overall
Use this checklist before and after an AI-assisted task
-
Confirm the tool and purpose are authorized
Check the agency’s current approved-tool list and whether authorization covers the particular task, account or configuration, and information class. If the policy does not answer those questions, ask the relevant agency official before proceeding.
-
Classify every prompt and upload
Treat anything entered into a system—including pasted text, uploaded files, and follow-up prompts—as information disclosed to that system. Do not submit personal information, confidential records, protected operational information, classified material, or other nonpublic agency information unless your agency expressly authorizes the tool for that information and use.
-
Understand data handling through your agency
Ask how prompts and outputs are retained, who can access them, whether they can be used for model training or other purposes, and how deletion works. Do not personally accept terms of service or create an official account unless agency rules allow it. Exact account and data-handling requirements vary, so use the agency’s process rather than deciding that a vendor’s terms are sufficient.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy. -
Check the output against authoritative sources
Verify factual claims, dates, laws, policies, citations, calculations, and any other material that matters to the task. Remove fabricated references and unsupported assertions, and check for biased or inappropriate content. Fluent wording is not proof that an answer is correct.
-
Review and own the work
Edit AI-assisted material before it becomes official work. You remain accountable for what you submit, communicate, or decide. Follow agency records rules for preserving material work or decision records, and do not let a system make—or effectively determine—a consequential government decision without agency authorization and required review.
-
Escalate uncertainty and recheck the rules
Consult agency privacy, security, legal, procurement, records, program, or AI governance staff if you are unsure about the information class, tool authorization, data treatment, effect on a person, or required oversight. Recheck current agency guidance as tools and policies change; a remembered rule or another agency’s practice may not apply.
Why data handling and contracts matter
Federal acquisition guidance makes privacy and government-data terms important when agencies obtain AI services. OMB Memorandum M-25-22 directs agencies to establish privacy processes for AI services handling personally identifiable information (PII), involve senior privacy officials early and continuously in relevant acquisitions, and set appropriate contractual terms for government data and intellectual property.
The memorandum says contracts must permanently prohibit use of nonpublic agency data supplied as input and results produced as output to further train publicly or commercially available AI algorithms unless the agency explicitly consents, consistent with applicable law. This is acquisition guidance, not a blanket instruction that every contract clause applies directly to each employee. In practice, employees should follow their agency’s implementation and not assume that a public-facing or consumer AI tool is authorized for official information. Read OMB Memorandum M-25-22.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Apply more scrutiny when AI affects a person
Routine drafting or brainstorming is different from using AI to inform hiring, eligibility, enforcement, or another decision affecting someone. Before using AI in a consequential process, pause and route the proposed use through the agency’s applicable legal, privacy, security, program, and AI governance review. Confirm what human review, risk assessment, monitoring, documentation, or other controls are required by the agency’s rules and the governing program.
Do not treat any one agency’s procedure as a universal legal requirement. OPM’s August 27, 2026 memorandum addresses AI in the federal hiring process and clarifies application of OMB Memorandum M-25-21 in that context; it is not a general policy for every agency AI use. GSA’s directive calls for assessment, evaluation, monitoring, and governance within its own defined scope. Read OPM’s federal hiring memorandum.
Why current agency guidance matters
AI use and policy are changing. In a July 29, 2025 report covering inventories from 11 selected federal agencies, GAO found that reported AI use cases rose from 571 in 2023 to 1,110 in 2024; reported generative AI cases rose from 32 to 282. These are agency-reported use cases in the selected agencies, not a census of government use or a measure of employee adoption.
Free tools Windows power users keep installed
One-click scans. No signup required.
GAO also reported that officials at 10 of 12 selected agencies said existing policy, including data-privacy policy, could present obstacles to adoption, while four said rapid technological change complicated policy and practice. The findings reinforce why employees should consult current agency instructions instead of assuming that approval, rules, or practices have stayed the same. Read GAO’s report.
How Canadian federal guidance differs
Government of Canada guidance is a separate jurisdictional example, not a U.S. federal rule. Its guide says public servants must not enter personal information in publicly available online generative AI tools, while use in government-controlled or configured systems may be appropriate when privacy and security controls are in place. It also distinguishes routine support, such as brainstorming or drafting ordinary correspondence, from uses that make or inform administrative decisions. For the latter, the Canadian guide points to that country’s Directive on Automated Decision-Making, including its Algorithmic Impact Assessment and related transparency, quality-assurance, and procedural-fairness requirements. Do not import those Canadian mechanisms into U.S. agency work. Read the Canadian guide.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
