diff --git a/Userland/Libraries/LibJS/Lexer.cpp b/Userland/Libraries/LibJS/Lexer.cpp
index 98d5efce98d..2a0b14d1baa 100644
--- a/Userland/Libraries/LibJS/Lexer.cpp
+++ b/Userland/Libraries/LibJS/Lexer.cpp
@@ -567,7 +567,7 @@ Token Lexer::next()
     // This is being used to communicate info about invalid tokens to the parser, which then
     // can turn that into more specific error messages - instead of us having to make up a
     // bunch of Invalid* tokens (bad numeric literals, unterminated comments etc.)
-    ByteString token_message;
+    StringView token_message;
 
     Optional<DeprecatedFlyString> identifier;
     size_t identifier_length = 0;
@@ -643,7 +643,7 @@ Token Lexer::next()
             m_parsed_identifiers->identifiers.set(*identifier);
         } else {
             token_type = TokenType::Invalid;
-            token_message = "Start of private name '#' but not followed by valid identifier";
+            token_message = "Start of private name '#' but not followed by valid identifier"sv;
         }
     } else if (auto code_point = is_identifier_start(identifier_length); code_point.has_value()) {
         bool has_escaped_character = false;
@@ -734,7 +734,7 @@ Token Lexer::next()
         }
         if (is_invalid_numeric_literal) {
             token_type = TokenType::Invalid;
-            token_message = "Invalid numeric literal";
+            token_message = "Invalid numeric literal"sv;
         }
     } else if (m_current_char == '"' || m_current_char == '\'') {
         char stop_char = m_current_char;
@@ -761,7 +761,7 @@ Token Lexer::next()
     } else if (m_eof) {
         if (unterminated_comment) {
             token_type = TokenType::Invalid;
-            token_message = "Unterminated multi-line comment";
+            token_message = "Unterminated multi-line comment"sv;
         } else {
             token_type = TokenType::Eof;
         }
@@ -844,7 +844,7 @@ Token Lexer::next()
     } else {
         m_current_token = Token(
             token_type,
-            String::from_byte_string(token_message).release_value_but_fixme_should_propagate_errors(),
+            token_message,
             m_source.substring_view(trivia_start - 1, value_start - trivia_start),
             m_source.substring_view(value_start - 1, m_position - value_start),
             m_filename,
diff --git a/Userland/Libraries/LibJS/Token.h b/Userland/Libraries/LibJS/Token.h
index 4058e4202b4..c3a91dfba48 100644
--- a/Userland/Libraries/LibJS/Token.h
+++ b/Userland/Libraries/LibJS/Token.h
@@ -181,9 +181,9 @@ class Token {
 public:
     Token() = default;
 
-    Token(TokenType type, String message, StringView trivia, StringView value, StringView filename, size_t line_number, size_t line_column, size_t offset)
+    Token(TokenType type, StringView message, StringView trivia, StringView value, StringView filename, size_t line_number, size_t line_column, size_t offset)
         : m_type(type)
-        , m_message(move(message))
+        , m_message(message)
         , m_trivia(trivia)
         , m_original_value(value)
         , m_value(value)
@@ -200,7 +200,7 @@ public:
     char const* name() const;
     static char const* name(TokenType);
 
-    String const& message() const { return m_message; }
+    StringView message() const { return m_message; }
     StringView trivia() const { return m_trivia; }
     StringView original_value() const { return m_original_value; }
     StringView value() const
@@ -246,7 +246,7 @@ public:
 
 private:
     TokenType m_type { TokenType::Invalid };
-    String m_message;
+    StringView m_message;
     StringView m_trivia;
     StringView m_original_value;
     Variant<Empty, StringView, DeprecatedFlyString> m_value {};