AI at Work Without Leaking the Company

Tóm tắt

The request arrives with no budget and no scope. It lands on the security lead, the IT lead, or on whoever in the company has actually read a set of terms all the way through. That is usually you. The first thing to accept is that you are not at the start. You are somewhere in the middle. People in your company have been pasting work into chat windows on personal accounts since long before anyone asked whether they should, and the question in front of you is not whether to allow it. It is what is already leaving, on which accounts, under which terms, and what you are going to do about it this month. Across four modules you will run a discovery that tells you what is actually in use rather than what you assume, work out precisely what leaves the building when somebody pastes, build an approved tool set short enough that people follow it, learn to test a vendor's retention and training claims instead of accepting them, sort your company's data into three classes concrete enough to print on one page, and decide which decisions in your organisation may never be handed to a machine — and how you would prove a human actually looked. The method for the policy itself is borrowed wholesale from the compliance segment, B1, and it is the same rule for the same reason: write the policy to the state you are actually in today, then improve it. An aspirational AI policy is a signed, dated description of your own non-compliance, and it is worse than having no policy at all. Segment A1, AI Without the Hype, covers how the tool works and the personal habit of redacting before you paste. This segment does not repeat either. It is the organisational version: the controls, the written evidence, and the accountability when work produced this way goes out under your company's name. **This is not legal advice.** Where UK GDPR or the EU AI Act touch what you are doing, this segment describes the direction those obligations point in and stops there. Your data protection officer and your lawyer decide the rest, and the sentence "our security lead read a course" has never once helped anybody in a regulatory conversation.

1. 01 It Is Already Happening

   3 bài học·1h 46m

   You do not have a decision to make about whether to allow this. That decision was made months ago by about forty people independently, none of whom told you. Your first job is to find out what they…

2. 02 Tools, Vendors and Written Controls

   3 bài học·1h 36m

   A ban list is a list of things to route around. Every name you add to it makes the list less likely to be read and more likely to be treated as a challenge. The control that works is the short list…

3. 03 What May Be Pasted, and What May Go Out

   3 bài học·1h 31m

   A rule that requires judgement at the moment of pasting is not a rule. It is a hope. This module produces a table somebody can read in four seconds while holding a customer's file.

4. 04 The Policy, the Human, and the First Hour

   3 bài học·2h 3m

   The rule is borrowed from B1 and it does not change because the subject is AI. Write the policy to the state you are actually in today. Then improve it, and change the page on the day the work lands.

Các câu hỏi thường gặp

Tôi sẽ học gì trong AI at Work Without Leaking the Company?

Somebody has handed you "sort out our AI policy". Your staff are already using these tools on personal accounts with company data, and have been for months. This segment gets you from that position to three artefacts you can put in front of a director: a one-page acceptable-use policy that is true t

Tôi có cần kinh nghiệm trước đó không?

Nên có một số kiến thức trước đó trước khi bắt đầu AI at Work Without Leaking the Company.

AI at Work Without Leaking the Company mất bao lâu?

AI at Work Without Leaking the Company bao gồm 4 mô-đun và 12 bài học. Bạn học theo tốc độ của riêng mình.

Làm cách nào để tôi có quyền truy cập?

AI at Work Without Leaking the Company được bao gồm trong bất kỳ gói đăng ký nào.