Washington Put AI on the Education Grant Scoreboard. The Privacy Rules Are Still in the Parking Lot.

I was camped out under the polite fluorescent hum of a public library when the Federal Register did what it always does: it reminded me that power rarely kicks down your door anymore. It mails you a notice. Then it funds you.

Education Department finalizes a grantmaking priority to advance AI in education

On April 13, 2026, the U.S. Department of Education published a final priority and definitions for its discretionary grant programs under 34 CFR Part 75 (docket ID ED-2025-OS-0118). The Department says the priority can be used across discretionary grant programs now and in the future, and it takes effect May 13, 2026. It also notes that more than 300 parties submitted comments on the proposed version first published in July 2025.

The core move is simple: if you want federal education dollars, it helps to speak Washington’s new dialect about AI literacy, educator training, and responsible adoption. The final text leans into:

• Age-appropriate approaches in K-12
• Training and support for educators
• Universal design for learning, so students with disabilities are not shoved to the margins
• Using AI technology to improve program outcomes (grant-speak for: show results, not just gadgets)

The Orwell check

Watch what happens when the word responsible does the job of a policy. Commenters urged the Department to require parental notification and opt-out provisions when AI tools are used in schools, and to mandate stronger privacy and safety requirements. The Department says it is committed to student privacy protections under law, but it declines to enact federal requirements, arguing safety and communication about technology are best decided at the state and local level.

The liberty ledger

Who gains: districts and administrators get a new, fundable lane for AI training and tools. Vendors get the familiar federal signal that turns pilots into platforms.

Who risks losing: students and families may get better instruction and support, but they also risk losing the practical ability to keep a child’s learning life from becoming a data exhaust pipe. The text acknowledges privacy worries raised by commenters, including parental consent and vendor disclosure concerns, then stops short of requiring those protections federally.

The tradeoff

The Department argues that global competitiveness means students need opportunities to learn to use AI effectively. Fine. But the tradeoff here is not innovation versus stagnation. It is innovation versus accountability, especially because this priority can ripple across many discretionary programs.

The Paine test

Does this expand liberty or concentrate power? Done right, AI literacy expands liberty by teaching students to understand, critique, and resist automated nonsense. The Department even revised its AI literacy definition to include ethical reasoning, critical social inquiry, interdisciplinary problem-solving, and creativity. Done wrong, AI in education concentrates power in a triangle: government money, vendor systems, and institutional convenience.

If Washington wants AI on the grant scoreboard, it should treat baseline safeguards like part of the equipment list: data minimization, clear retention limits, public disclosure of tools used, independent evaluation of impacts, and meaningful opt-out paths that do not punish students.

Accountability is available if anyone feels like using it: Congress can ask oversight questions and require reporting conditions; the Department’s Inspector General can audit procurement and data-protection practices tied to these priorities; states can set uniform student data rules; and school boards can demand vendor transparency before signing anything that touches a child account, learning profile, or behavioral log.

Washington has a May 13, 2026 effective date and a nationwide incentive tailwind. If we are going to subsidize this future, why are parental notice, opt-out rights, and public audits still treated like optional accessories?