diff --git a/code/modules/integrated_electronics/subtypes/input.dm b/code/modules/integrated_electronics/subtypes/input.dm
index 81b4a67f3b..c0ac9817fa 100644
--- a/code/modules/integrated_electronics/subtypes/input.dm
+++ b/code/modules/integrated_electronics/subtypes/input.dm
@@ -526,6 +526,7 @@
 	complexity = 8
 	inputs = list()
 	outputs = list(
+	"speaker ref",//CHOMPADDITION: so we can target the speaker with an action
 	"speaker" = IC_PINTYPE_STRING,
 	"message" = IC_PINTYPE_STRING
 	)
@@ -551,8 +552,9 @@
 			// as a translation, when it is not.
 			if(S.speaking && !istype(S.speaking, /datum/language/common))
 				translated = TRUE
-		set_pin_data(IC_OUTPUT, 1, M.GetVoice())
-		set_pin_data(IC_OUTPUT, 2, msg)
+		set_pin_data(IC_OUTPUT , 1, weakref(M))//CHOMPADDITION: so we can target the speaker with an action
+		set_pin_data(IC_OUTPUT, 2, M.GetVoice())
+		set_pin_data(IC_OUTPUT, 3, msg)
 
 	push_data()
 	activate_pin(1)
@@ -570,6 +572,7 @@
 	complexity = 12
 	inputs = list()
 	outputs = list(
+	"speaker ref",//CHOMPADDITION: so we can target the speaker with an action
 	"speaker" = IC_PINTYPE_STRING,
 	"message" = IC_PINTYPE_STRING
 	)
@@ -602,8 +605,9 @@
 		for(var/datum/multilingual_say_piece/S in message_pieces)
 			if(!((S.speaking.flags & NONVERBAL) || (S.speaking.flags & SIGNLANG))||S.speaking.name == LANGUAGE_ECHOSONG) //Ignore verbal languages
 				return
-		set_pin_data(IC_OUTPUT, 1, M.GetVoice())
-		set_pin_data(IC_OUTPUT, 2, msg)
+		set_pin_data(IC_OUTPUT , 1, weakref(M))//CHOMPADDITION: so we can target the speaker with an action
+		set_pin_data(IC_OUTPUT, 2, M.GetVoice())
+		set_pin_data(IC_OUTPUT, 3, msg)
 
 	push_data()
 	if(!translated)