When dealing with a digital service, it is often necessary to compare hashes manually to certify the integrity of data. This study evaluates two alternative encodings - emoji sequences and word sequences - to improve the usability of integrity checks. We conducted a web-based user study to assess participants’ ability to recognize attacks, collecting insights from both a pilot test and then a main test. Results show that users achieved over 70% accuracy in anomaly detection, demonstrating the potential of these encodings to enhance verifiability while maintaining usability.