{"id":793,"date":"2021-12-06T09:23:35","date_gmt":"2021-12-06T09:23:35","guid":{"rendered":"https:\/\/www.workfall.com\/learning\/blog\/?p=793"},"modified":"2025-08-20T11:09:11","modified_gmt":"2025-08-20T11:09:11","slug":"how-to-add-privacy-to-your-transcriptions-with-amazon-transcribe-to-hide-all-the-confidential-information-from-the-audio-fetched-text-part-2","status":"publish","type":"post","link":"https:\/\/learning.workfall.com\/learning\/blog\/how-to-add-privacy-to-your-transcriptions-with-amazon-transcribe-to-hide-all-the-confidential-information-from-the-audio-fetched-text-part-2\/","title":{"rendered":"How to add privacy to your transcriptions with Amazon Transcribe to hide all the confidential information from the audio-fetched text? (Part 2)"},"content":{"rendered":"Reading Time: <\/span> 10<\/span> minutes<\/span><\/span>\n

<\/p>\n\n\n\n

Using Amazon Transcribe, AWS customers can easily add speech-to-text capabilities to their applications via an automatic speech recognition service. In short, Amazon Transcribe helps in redacting all the Personally Identifiable Information (PII) instead of doing it manually which is a very tedious effort. A transcription job helps to analyze the audio files that contain voice into text with the use of advanced machine learning techniques to perform the conversion. The automatic redaction feature (that is enabled while creating a new transcription job) redacts all the personally identifiable information from the output results and replaces any such information with a [PII] tag in the output result. The Amazon Transcribe service can be integrated with multiple other services. We can also make use of Amazon Lambda<\/a> to create a transcription job and trigger the job for a file in an S3 bucket<\/a>. In this blog, we will learn how to add privacy to the transcriptions via the content removal redaction feature of Amazon Transcribe to hide the confidential information in the fetched text from an audio file.<\/p>\n\n\n\n

Refer to part 1 of the blog here<\/a>.<\/p>\n\n\n\n

Amazon Transcribe – Recap<\/h2>\n\n\n\n
\"Amazon<\/figure>\n\n\n\n

Amazon Transcribe<\/a> makes converting speech to text simple for customers. It can be used for a variety of business applications, including transcription of voice-based customer service calls, generation of subtitles on audio\/video content, and conducting (text-based) content analysis on audio\/video content, using Automatic Speech Recognition (ASR) technology.<\/p>\n\n\n\n

Any device having an on-device microphone, such as phones, PCs, tablets, and IoT<\/a> devices, can use Amazon Transcribe. The Amazon Transcribe API will be able to detect the quality of the audio stream being input to the device and will choose the appropriate acoustic models for speech-to-text conversion accordingly. Additionally, developers can use the Transcribe API to access speech-to-text conversion capabilities in their applications.<\/p>\n\n\n\n

Hands-on<\/h2>\n\n\n\n
\"Amazon<\/figure>\n\n\n\n

In this blog, we will see how we can add privacy to transcriptions with Amazon Transcribe. The service will in-turn take an audio file as an input and provide us with the text in the audio with all the confidential information hidden. Navigating to the provided link, we will first download the sample audio file. We will then navigate to the Amazon S3 dashboard and create a new S3 bucket. After the creation of a new bucket, we will upload the downloaded sample audio file to the bucket with bucket versioning enabled. Navigating to the uploaded object dashboard, we will copy its S3 URI. Navigating to the Amazon Transcribe dashboard, we will create a new transcription job with content removal enabled for the job and with the input data as the S3 URI of the newly added object in the bucket. We will then traverse to the created job dashboard to test out the working and to see if the service has hidden all the confidential information from the text fetched from the audio file. Downloading the unredacted and the redacted JSON files, we will note down the difference. We will then navigate to the Amazon Lambda dashboard and create a new Lambda function with the runtime as Python 3.9. Copying the provided Amazon Transcribe API code and pasting it into the functions code editor pane, we will create a new test event. The pasted code will in-turn create a new transcription job, fetch the audio file from the provided S3 URI, pass it to the transcribe job and then store the results in the form of JSON outputs in the provided S3 buckets.  We will then navigate to the Lambda function configuration IAM role portal to attach the Amazon S3 and Amazon Transcribe services required policies. Refreshing the lambda function page, we will test out the code. We will then navigate to the Transcription jobs dashboard and check if the lambda function successfully created a new job. Then we will navigate to the Amazon S3 bucket to check if the function stored the required output files in the bucket. Finally, we will download both the output files and review it with the previous jobs output files.<\/p>\n\n\n\n

To implement this, we will do the following:<\/strong><\/p>\n\n\n\n