Please enable JavaScript.
Coggle requires JavaScript to display documents.
AI and ML, Al/ML can strengthen cybersecurity by helping detect, prevent,…
AI and ML
Cybersecurity Advantages
-
-
-
-
Ethical Theory—Utilitarianism: choose the option that provides the greatest overall good by reducing harm for the most people
Scenario 1: Detecting Unusual Behavior
An AI system notices unusual login patterns that might indicate a compromised account. :star:
Scenario 2: Automating Routine Tasks
A security team is overwhelmed with routine alerts, such as spam or low-risk events. :star:
-
-
-
-
Ethical Theory
Virtue ethics - using technology responsibly shows good judgment and helps the team work more effectively.
Ethical Challenges
-
Option B
Audit the system, use diverse data, and keep human review in the process.
Scenario 1: AI in Hiring
A company uses Al to screen job applicants, but there are concerns the system may be biased against certain groups. :star:
-
-
Ethical Theory - Justice (John Rawls) - decisions should be fair and give equal opportunity to everyone.
Scenario 2: Job Displacement
A company plans to replace some customer service roles with an Al chatbot. :star:
-
-
-
-
Ethical Theory
Care ethics - consider the well-being of people and the relationships affected by the decision.
-
Cybersecurity Risks
Ethical Method: Precautionary principle - Take reasonable steps to prevent harm when there is a known risk.
-
-
Scenario 2: Al for Cyberattacks
Generative Al tools can be used to create more convincing phishing emails and other attacks. :star:
-
-
-
-
-
-
Scenario 1: Adversarial Attacks
Attackers try to manipulate an AI system by feeding it specially crafted inputs to make it misclassify a threat. :star:
-
Al/ML can strengthen cybersecurity by helping detect, prevent, and respond to threats more effectively.
AI/ML raise ethical questions related to fairness, transparency, accountability, job displacement, and the impact on society.
AI/ML rely on large amounts of data, which can create privacy concerns and new security vulnerabilities.
AI/ML can introduce new risks, such as adversarial attacks, biased outputs, data poisoning, and the potential for misuse by malicious actors.