diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index b0c1c98ef9..be3f710f3d 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -287,7 +287,7 @@ jobs:
           platform: "android"
           runs-on: "ubuntu-latest"
           briefcase-run-prefix: JAVA_HOME=${JAVA_HOME_17_X64}
-          briefcase-run-args: >
+          briefcase-run-args: >-
             --device '{"avd":"beePhone","skin":"pixel_3a"}'
             --Xemulator=-no-window
             --Xemulator=-no-snapshot
@@ -333,7 +333,7 @@ jobs:
       timeout-minutes: 15
       run: |
         ${{ matrix.briefcase-run-prefix }} \
-          briefcase run ${{ matrix.platform }} --log --test ${{ matrix.briefcase-run-args }}
+          briefcase run ${{ matrix.platform }} --log --test ${{ matrix.briefcase-run-args }} -- --ci
 
     - name: Upload Logs
       uses: actions/upload-artifact@v4.3.3
diff --git a/testbed/tests/testbed.py b/testbed/tests/testbed.py
index 00a6858ead..de96b7137b 100644
--- a/testbed/tests/testbed.py
+++ b/testbed/tests/testbed.py
@@ -13,7 +13,7 @@
 from testbed.app import main
 
 
-def run_tests(app, cov, args, report_coverage, run_slow):
+def run_tests(app, cov, args, report_coverage, run_slow, running_in_ci):
     try:
         # Wait for the app's main window to be visible.
         print("Waiting for app to be ready for testing... ", end="", flush=True)
@@ -26,6 +26,8 @@ def run_tests(app, cov, args, report_coverage, run_slow):
         project_path = Path(__file__).parent.parent
         os.chdir(project_path)
 
+        os.environ["RUNNING_IN_CI"] = "true" if running_in_ci else ""
+
         app.returncode = pytest.main(
             [
                 # Output formatting
@@ -146,6 +148,12 @@ def run_tests(app, cov, args, report_coverage, run_slow):
     except ValueError:
         report_coverage = False
 
+    try:
+        args.remove("--ci")
+        running_in_ci = True
+    except ValueError:
+        running_in_ci = False
+
     # If there are no other specified arguments, default to running the whole suite,
     # and reporting coverage.
     if len(args) == 0:
@@ -160,6 +168,7 @@ def run_tests(app, cov, args, report_coverage, run_slow):
             args=args,
             run_slow=run_slow,
             report_coverage=report_coverage,
+            running_in_ci=running_in_ci,
         )
     )
     # Queue a background task to run that will start the main thread. We do this,
diff --git a/testbed/tests/widgets/test_canvas.py b/testbed/tests/widgets/test_canvas.py
index e351ed1364..3ac2878766 100644
--- a/testbed/tests/widgets/test_canvas.py
+++ b/testbed/tests/widgets/test_canvas.py
@@ -1,4 +1,5 @@
 import math
+import os
 from math import pi, radians
 from unittest.mock import Mock, call
 
@@ -693,6 +694,10 @@ async def test_write_text(canvas, probe):
     assert_reference(probe, "write_text", threshold=0.07)
 
 
+@pytest.mark.xfail(
+    condition=os.environ.get("RUNNING_IN_CI") != "true",
+    reason="may fail outside of CI",
+)
 async def test_multiline_text(canvas, probe):
     "Multiline text can be measured and written"