Quiz połączony z przewodnikiem · Średni Poziom
Napisanie quizu dotyczącego zasad dopuszczalnego użytkowania sztucznej inteligencji w miejscu pracy
Omawia zatwierdzone poziomy narzędzi, mapowanie klas danych do narzędzi, zasady przeglądu ręcznego i ujawniania informacji oraz to, dlaczego ogólne zakazy dotyczące sztucznej inteligencji zwykle przynoszą odwrotny skutek.
Pytanie 1 z 8
Why does the guide recommend keeping the approved tools list as a separate living document?
Podgląd każdego pytania8 pytania
Odpowiedzi i wyjaśnienia pojawiają się w miarę rozwiązywania quizu.
- Why does the guide recommend keeping the approved tools list as a separate living document?
- Regulators require a separate file
- Tools change often, so the list can be updated without rewriting the policy
- It hides the list from employees
- Separate documents cannot be edited
- How should a policy define which information may go into which tools?
- By linking to the organization's existing data classification levels
- By letting each employee decide case by case
- By banning all data from every tool
- By relying on the vendor's marketing claims
- What does the guide say tends to happen with a blanket ban on AI tools?
- It removes all AI risk
- It speeds up tool procurement
- It improves employee training results
- It pushes use onto personal devices where nobody can see it
- Under the human review section, who is responsible for the accuracy of AI output?
- Dostawca sztucznej inteligencji
- Dział IT
- Osoba korzystająca z danych wyjściowych
- Nobody, if the tool was approved
- Which feature fits a Tier A tool in the guide's matrix?
- A free consumer account with default settings
- Contract terms excluding training on your data, plus single sign-on and audit logs
- Any tool an employee installs personally
- A tool with no retention controls
- Which framework did NIST publish in January 2023 that can inform an AI policy?
- The AI Risk Management Framework
- ISO/IEC 42001
- The General Data Protection Regulation
- The SOC 2 standard
- Why should the policy's scope include AI features built into software the company already uses?
- Built-in features never process data
- Built-in features are always illegal
- Only built-in features need disclosure
- AI features often arrive through routine updates to existing software
- Which item belongs on a vendor review checklist?
- The vendor's office decor
- The number of social media followers
- Data retention periods and whether inputs are used for training
- Kolor interfejsu użytkownika