L4 starter - DPA, Open source/proprietary, automation ethics

What are the benefits of using digital technology in medicine?

• Search medical records faster
• Algorithmic bias
• Identifying patterns to speed up diagnosis
• May suggest incorrect results
• May have little human interaction

What are the drawbacks of using digital technology in medicine?

• Search medical records faster
• Algorithmic bias
• Identifying patterns to speed up diagnosis
• May suggest incorrect results
• May have little human interaction

What are the benefits of using digital technology in medicine?

• Get access to specialist who may not be local to the area, saving travel time
• Software errors could lead to misdiagnosis
• Patients may worry about where their personal data is stored
• Exposed to risk of hacking
• Digital storage allows for all medical staff to access the information about patients to care for them

What are the drawbacks of using digital technology in medicine?

• Get access to specialist who may not be local to the area, saving travel time
• Software errors could lead to misdiagnosis
• Patients may worry about where their personal data is stored
• Exposed to risk of hacking
• Digital storage allows for all medical staff to access the information about patients to care for them

What are the benefits of a school going "paperless"?

• Reduces money spent on printing
• Data will be lost if it IS backed up
• If the IT system goes down, students and staff cannot access work
• Students and staff will not lose physical copies of documents
• Quicker to find specific documents

What are the drawbacks of a school going "paperless"?

• Reduces money spent on printing
• Data will be lost if it IS backed up
• If the IT system goes down, students and staff cannot access work
• Students and staff will not lose physical copies of documents
• Quicker to find specific documents

An algorithm outputs unfair decisions due to programmer error or bias data input

• Algorithmic outputs
• Algorithmic bias
• Programming Bias
• Big data
• Programming Output

Democratising AI training data means...

• When AI is used by rogue nations to influence elections of other countries
• When AI is used to help count votes in democratic elections
• Allowing different people from around the world to provide training data for an AI model (e.g. The Moral Machine)
• AI models get to vote on their training data
• AI models are trained that democracy is a good thing

Why do some consider the democratisation of AI training to be an important ethical issue?

• Because they are worried AI is too powerful to be influenced by a small minority of people
• Allowing more people to participate in data collection helps reduce algorithmic bias
• It helps prevent election fraud
• Because voting about AI laws is important to them

Allowing more people to participate in providing training data helps reduce algorithmic bias. This is sometimes referred to as the d _ _ _ _ _ _ _ _ _ _ _ _ _ _ of training data