WEBVTT 00:00:00.000 --> 00:00:13.274 **"Every time you paste your resume into an AI tool, you might be training 00:00:13.374 --> 00:00:26.648 someone else's model with your personal data."** 00:00:26.748 --> 00:00:40.023 Welcome to Module 5, Lesson 5. Today, we are talking about Data Privacy and AI 00:00:40.123 --> 00:00:53.397 Hygiene. As you leverage Generative AI to speed up your job search, write cover 00:00:53.497 --> 00:01:06.772 letters, and draft case studies, you are interacting with powerful systems that 00:01:06.872 --> 00:01:20.146 ingest data. Most job seekers use AI like a standard search engine, but a career 00:01:20.246 --> 00:01:33.521 strategist uses it with extreme discipline. 00:01:33.621 --> 00:01:46.895 If you don't understand the boundaries of information sharing, you risk 00:01:46.995 --> 00:02:00.269 compromising your personal privacy, violating legal obligations like NDAs, and 00:02:00.369 --> 00:02:13.644 permanently damaging your professional integrity. Today, I'm going to walk you 00:02:13.744 --> 00:02:27.018 through the exact protocol for protecting yourself. We’ll cover the 5-Point 00:02:27.118 --> 00:02:40.393 Privacy Audit, the Data Hygiene Protocol, platform-specific rules, and the 00:02:40.493 --> 00:02:53.767 "Burner Profile" strategy. 00:02:53.867 --> 00:03:07.142 Let's define our core term: AI hygiene. This is the disciplined practice of 00:03:07.242 --> 00:03:20.516 managing the data you provide to artificial intelligence models to prevent the 00:03:20.616 --> 00:03:33.891 leakage of sensitive, confidential, or proprietary information. 00:03:33.991 --> 00:03:47.265 When you ask an AI to help you build a resume or prep for an interview, you are 00:03:47.365 --> 00:04:00.639 feeding it context. If you feed it raw internal company metrics or proprietary 00:04:00.739 --> 00:04:14.014 software code, you are violating professional ethics. **Sharing raw metrics can 00:04:14.114 --> 00:04:27.388 violate Non-Disclosure Agreements (NDAs) and expose a former employer's 00:04:27.488 --> 00:04:40.763 non-public performance data**. If that proprietary information leaks through an 00:04:40.863 --> 00:04:54.137 AI model's training data, the consequences for your reputation—and your legal 00:04:54.237 --> 00:05:07.512 standing—can be severe. 00:05:07.612 --> 00:05:20.886 Protecting your previous employer's intellectual property isn't just a legal 00:05:20.986 --> 00:05:34.260 obligation; it's a demonstration of your professional integrity to your future 00:05:34.360 --> 00:05:47.635 employer. 00:05:47.735 --> 00:06:01.009 To use Generative AI effectively without compromising security, you must master 00:06:01.109 --> 00:06:14.384 the art of data anonymization. My Data Hygiene Protocol dictates exactly what 00:06:14.484 --> 00:06:27.758 you can share, and what is absolutely prohibited. 00:06:27.858 --> 00:06:41.133 First, let's cover the Data Prohibitions. There are four categories of 00:06:41.233 --> 00:06:54.507 information you must never paste into a GenAI prompt: 00:06:54.607 --> 00:07:07.882 1. Personally Identifiable Information (PII): Specific personal details that 00:07:07.982 --> 00:07:21.256 could link you or a colleague to a dataset. 00:07:21.356 --> 00:07:34.630 2. Proprietary Information: Internal processes, trade secrets, or software 00:07:34.730 --> 00:07:48.005 code owned by your current or former employer. 00:07:48.105 --> 00:08:01.379 3. Financial and Strategic Data: Non-public revenue figures, internal growth 00:08:01.479 --> 00:08:14.754 metrics, or strategic company roadmaps. 00:08:14.854 --> 00:08:28.128 4. Contractual Data: Any information strictly governed by an NDA. 00:08:28.228 --> 00:08:41.503 So, how do you get AI to help you write a tailored resume or case study without 00:08:41.603 --> 00:08:54.877 sharing this data? You use Anonymization Strategies to strip away specific 00:08:54.977 --> 00:09:08.251 identifiers while keeping the structural value of the data. 00:09:08.351 --> 00:09:21.626 Here are the four techniques you need to use: 00:09:21.726 --> 00:09:35.000 * Generalization: Replace specific company names with industry descriptors. 00:09:35.100 --> 00:09:48.375 Instead of saying "Company X," write "A Top-5 FinTech firm". 00:09:48.475 --> 00:10:01.749 * Data Masking: Substitute actual financial figures or user counts with 00:10:01.849 --> 00:10:15.124 percentages. Write "Increased efficiency by 20%" instead of pasting raw internal 00:10:15.224 --> 00:10:28.498 logs. 00:10:28.598 --> 00:10:41.873 * Abstraction: Remove specific technical architecture details and replace 00:10:41.973 --> 00:10:55.247 them with high-level conceptual frameworks. This allows you to demonstrate 00:10:55.347 --> 00:11:08.621 technical proficiency without revealing proprietary implementation details. 00:11:08.721 --> 00:11:21.996 * Redaction: Completely remove any mention of specific projects, product 00:11:22.096 --> 00:11:35.370 codenames, or stakeholder names. 00:11:35.470 --> 00:11:48.745 When building case studies, always prioritize "Framework over Content." Ask 00:11:48.845 --> 00:12:02.119 the AI to help you structure the narrative or improve the flow, but only provide 00:12:02.219 --> 00:12:15.494 it with a generalized problem-solution framework, or transform real-world 00:12:15.594 --> 00:12:28.868 projects into hypothetical scenarios. 00:12:28.968 --> 00:12:42.243 Before you type a single word into a new AI tool, you need to run it through the 00:12:42.343 --> 00:12:55.617 5-Point Privacy Audit. This is your checklist for evaluating any AI tool 00:12:55.717 --> 00:13:08.991 before using it. 00:13:09.091 --> 00:13:22.366 1. Check the Training Policy: Does the platform use your input data to train 00:13:22.466 --> 00:13:35.740 their public models? 00:13:35.840 --> 00:13:49.115 2. Opt-Out Capability: Is there a clear, accessible toggle to turn off data 00:13:49.215 --> 00:14:02.489 sharing and model training? 00:14:02.589 --> 00:14:15.864 3. Data Retention: How long does the platform store your chat logs and 00:14:15.964 --> 00:14:29.238 uploaded documents on their servers? 00:14:29.338 --> 00:14:42.612 4. Enterprise vs. Consumer Tiers: Are you using a free consumer tier (which 00:14:42.712 --> 00:14:55.987 often harvests data) or a secure enterprise/paid tier that guarantees privacy? 00:14:56.087 --> 00:15:09.361 5. Output Risk Assessment: If this prompt were accidentally made public, 00:15:09.461 --> 00:15:22.736 would it breach an NDA or expose PII? 00:15:22.836 --> 00:15:36.110 If a tool fails this audit, do not feed it your career history. 00:15:36.210 --> 00:15:49.485 Not all AI tools treat your data the same way. You need to know **which tools 00:15:49.585 --> 00:16:02.859 store your data, which don't, and how to check**. 00:16:02.959 --> 00:16:16.234 As part of your job search operations, make it a habit to navigate directly to 00:16:16.334 --> 00:16:29.608 the privacy settings of whatever major AI platform you are using. Look for the 00:16:29.708 --> 00:16:42.982 "Data Controls" or "Privacy" tab. Most major AI platforms have a default setting 00:16:43.082 --> 00:16:56.357 that opts you into model training. You must manually go in and toggle that 00:16:56.457 --> 00:17:09.731 setting off. Remember, your resume is a goldmine of data—don't hand it over for 00:17:09.831 --> 00:17:23.106 free just because you skipped the settings menu. 00:17:23.206 --> 00:17:36.480 Finally, I want to introduce you to a safety mechanism called the **"Burner 00:17:36.580 --> 00:17:49.855 Profile" strategy**. 00:17:49.955 --> 00:18:03.229 When you are testing a new AI platform, optimizing a prompt, or experimenting 00:18:03.329 --> 00:18:16.603 with the AI Prompt Library, you shouldn't use your real resume. Instead, use a 00:18:16.703 --> 00:18:29.978 Burner Profile—a document using **modified personal info for AI testing before 00:18:30.078 --> 00:18:43.352 committing real data**. 00:18:43.452 --> 00:18:56.727 Change your name, alter your exact job titles slightly, shift your dates of 00:18:56.827 --> 00:19:10.101 employment, and use the anonymization techniques we discussed earlier to mask 00:19:10.201 --> 00:19:23.476 your companies. You run this Burner Profile through the AI to test the quality 00:19:23.576 --> 00:19:36.850 of its output, check how it behaves, and refine your prompts. Once you trust the 00:19:36.950 --> 00:19:50.225 prompt structure and have confirmed your privacy settings are locked down, only 00:19:50.325 --> 00:20:03.599 then do you apply it carefully to your actual documents. 00:20:03.699 --> 00:20:16.973 Your job search requires leverage, and AI provides that leverage—but only if you 00:20:17.073 --> 00:20:30.348 protect your downside. 00:20:30.448 --> 00:20:43.722 Your immediate implementation step for today: Download the **Personal AI 00:20:43.822 --> 00:20:57.097 Privacy Audit worksheet** attached to this lesson. I want you to run the 00:20:57.197 --> 00:21:10.471 checklist on every AI tool you currently use. Paste any document you plan to 00:21:10.571 --> 00:21:23.846 share with an AI into your own mental filter first to assess the privacy risk. 00:21:23.946 --> 00:21:37.220 Protect your data, protect your reputation, and I'll see you in the next lesson.