Ikibazo gihuza ibibazo · Biroroshye Urwego
Kubaka Ikarita yo Kwemeza AI
Ikizamini cyerekana ikarita yicyiciro, uburyo kwishingikiriza kumarembo yimiyoborere bishyiraho gahunda, nuruhare rwibikorwa nka NIST AI RMF na ISO / IEC 42001.
Ikibazo 1 Bya 8
According to the guide, why do many AI efforts stall?
Reba buri kibazo8 ibibazo
Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.
- According to the guide, why do many AI efforts stall?
- There are too few possible use cases
- Pilots never get the data, governance, budget or ownership needed to reach production
- AI models are too expensive to test
- Employees refuse to try new tools
- Why should a pilot's success criteria be set before it starts?
- So the vendor can price the pilot
- So the pilot can skip governance review
- So staff do not need training
- So the decision to scale is judged against an agreed baseline rather than chosen afterward
- What are the four functions of the NIST AI Risk Management Framework?
- Imiyoborere, Ikarita, Igipimo, Gucunga
- Tegura, Wubake, Ikizamini, Kohereza
- Discover, Pilot, Scale, Retire
- Menya, Kurinda, Gutahura, Gusubiza
- What is a governance gate in an AI roadmap?
- A firewall rule that blocks AI websites
- A budget approval for buying hardware
- A checkpoint with defined criteria that must be met before a use case moves to the next stage
- A meeting where vendors pitch products
- Why does the manufacturer schedule predictive maintenance after its sensor-data integration project?
- Predictive maintenance is illegal without integration
- The model depends on machine data flowing into one place, which the integration provides
- Integration projects are always cheaper
- The vendor required that order
- What is ISO/IEC 42001?
- A cybersecurity encryption protocol
- A law banning certain AI systems
- A programming standard for neural networks
- A management system standard for AI
- Which trait best describes a good early quick win?
- Low risk, visible, and possible with data that already exists
- The most technically ambitious project available
- A customer-facing system with no human review
- A project that needs a new data platform first
- In a hub-and-spoke AI operating model, what does the central hub usually do?
- Builds and owns every use case itself
- Only negotiates software discounts
- Sets standards, platforms and guardrails while business units build and own use cases
- Replaces the IT department