Information security risks: Is AI a threat to privacy?

1. Information disclosure/privacy: How can we prevent information from being leaked?

1.1. Technical solution

1.1.1. Companies must adopt technical approaches.

1.1.1.1. Federated learning

1.1.1.1.1. It allows data scientists to train shared statistical models using decentralised devices or local data sets.

1.1.1.2. Homomorphic encryption

1.1.1.2.1. It allows you to evaluate and manipulate encrypted data without disclosing it to anyone.

1.1.1.3. Differential privacy

1.1.1.3.1. The query result cannot be used to infer much about any specific individual if the effect of making an arbitrary single substitution in the database is modest enough, providing anonymity.

1.2. Behavioural Solution

1.2.1. Patch or update the software at any time.

1.2.1.1. Updating virus and Trojan libraries to prevent your computer from being compromised by new attacks.

1.2.2. Enforce authentication and encourage users to use password managers.

1.2.2.1. Encourage users to use multiple different passwords.

1.2.2.2. Give feedback to users if there is a risk of information leakage, and guide and help them to change their passwords.

1.2.3. Employees need to learn how to avoid attacks and security practices.

1.2.3.1. Reducing the potential for information crime through quality education for students.