NSF awarded UT Austin a grant to replace hand written robot safety rules with a model trained on demonstrated safe behavior
The National Science Foundation made an award to the University of Texas at Austin on August 13th, 2026 under its Foundational Research in Robotics program. The project trains a model on expert demonstrations and uses it as a filter that pulls a robot's action back toward behavior that was demonstrated as safe. The award record states the obligation as FY 2026 = $360,000.00 and runs the project from October 1st, 2026 through September 30th, 2029. Its own abstract is the argument for it, because it says engineers today must define every dangerous situation by hand.
How to take part
The award record is public and keyless at the National Science Foundation award service, where the abstract, the obligation, the program and the principal investigator are all published. The awardee institution is a public university and its research office is reachable.
Where
1 county. In Austin-Round Rock-San Marcos.
Timeline
- ordered
Award date recorded in the National Science Foundation award record
- Today
How this decision moved
One dated line per check, oldest first. A line that says nothing changed means somebody looked and it had not.
- 2026-08-21
Admitted. The award record names the University of Texas at Austin, states the obligation and dates the award to August 13th, 2026.
- 2026-08-25
Checked and unchanged. The decision still stands as decided.
- 2026-08-28
Checked and unchanged. The UT Austin robotics safety award still stands on the agency's own record with its stated obligation and project period unaltered.
- 2026-09-01
UT Austin's award for learning robot-safety constraints from demonstrations remains in the federal record.
- 2026-09-02
Checked and unchanged. The decision still stands as decided.
- 2026-09-05
The Austin award on learning robot safety rules from demonstrated behavior still stands. Nothing about the grant has moved.
The evidence
Every fact above rests on one of these. The words are the source's own.
"awardeeName":"University of Texas at Austin"National Science Foundation, award 2535276 Primary source, official · api.nsf.gov
For a robot to work safely alongside people, engineers today must manually define every dangerous situation the robot might encounter and write rules to avoid each one.National Science Foundation, award 2535276, abstract Primary source, official · api.nsf.gov
Existing safety methods in robotics require explicit mathematical descriptions of which states are safe, but these descriptions are difficult to specify, computationally expensive to enforce, and brittle in novel situations.National Science Foundation, award 2535276, abstract Primary source, official · api.nsf.gov
"fundsObligated":["FY 2026 = $360,000.00"]National Science Foundation, award 2535276 Primary source, official · api.nsf.gov
Questions about this decision
Answered from the record itself. Every answer is assembled from stored fields, so an answer the record has no basis for is left out rather than guessed.
What is this decision?
The National Science Foundation made an award to the University of Texas at Austin on August 13th, 2026 under its Foundational Research in Robotics program. The project trains a model on expert demonstrations and uses it as a filter that pulls a robot's action back toward behavior that was demonstrated as safe. The award record states the obligation as FY 2026 = $360,000.00 and runs the project from October 1st, 2026 through September 30th, 2029. Its own abstract is the argument for it, because it says engineers today must define every dangerous situation by hand.
Who decides it?
U.S. National Science Foundation decides. The record names the deciding body for every entry it carries.
Can the public take part?
The award record is public and keyless at the National Science Foundation award service, where the abstract, the obligation, the program and the principal investigator are all published. The awardee institution is a public university and its research office is reachable. No dated public window is on the record. The deciding body is named and reachable.
Where in Texas does it apply?
It covers Travis County in the Austin-Round Rock-San Marcos, TX area.
Has it been decided?
It has been decided. The dates on the item page carry when.
What happens next?
No future date is on the record. The last dated step on it was ordered on August 13th.
When did it start?
The earliest date on its record is August 13th, 2026.
What kind of decision is it?
It is filed under research and science.
What sources back it?
One source backs it. It is primary.
Is it on the ERCOT grid?
Yes. It sits inside the ERCOT interconnection.
When was it last checked?
Every fact on it was last verified against its source on September 5th, 2026.
Cite this
Texas AI Docket, NSF awarded UT Austin a grant to replace hand written robot safety rules with a model trained on demonstrated safe behavior. Tracked since August 13th, 2026. Last verified September 5th, 2026. https://texasaidocket.com/item/tx-2026-0081/. Reuse permitted under CC BY 4.0 with attribution. The same entry is in the docket JSON as item tx-2026-0081.
Beat
Filed under Research and science, with every other decision on that beat.