AI-GP

• Carl E. Mathis: 'We as humans are creating our AI algorithms, so therefore our AI algorithms inherit any biases that we might have.'
• Artificial intelligence (AI)


  Machines performing tasks that normally require human intelligence
• AI


  • A branch of computer science concerned with creating technology to do things that normally require human intelligence
  • Alan Turing developed a test to determine whether a machine is intelligent (1950)
  • A machine was considered intelligent if it produces responses to human interviewer that fool the interviewer into thinking the responses are human
• Definitions of AI


  Common elements: technology, autonomy, human involvement and output
• Technology


  Use of technology and specified objectives for the technology to achieve
• Autonomy


  Level of autonomy by the technology to achieve defined objectives
• Human involvement


  Need for human input to train the technology and identify objectives for it to follow
• Output


  Technology produces output, e.g., performing tasks, solving problems, producing content
• Machine learning


  The process of training machines to display AI behavior
• Types of machine learning

  • Supervised learning
  • Unsupervised learning
  • Reinforcement learning
• Supervised learning


  Labeled data that is grouped or classified into categories via the AI system
• Supervised learning


  • Text recognition, detecting spam in email
• Unsupervised learning


  Unlabeled data; typically used for pattern detection
• Unsupervised learning


  • Outliers in banking data; reviewing transactions for any fraudulent behavior
• Reinforcement learning


  An AI system is rewarded for performing a task well and penalized for not performing it well. Over time, learning to maximize the rewards and develop a system that works.
• Reinforcement learning


  • Self-driving cars: the system is rewarded when it keeps a car on the road and gets it to the destination where it is supposed to go. It is penalized if the car goes off the road or hits another object.
• AI as a socio-technical system

  AI influences society and society influences AI: strong relationship between the technology and human beings
• Relevant stakeholders to consider when working with AI

  • Individuals who look at the broader societal influences of AI, such as anthropologists, sociologists or others who work in social sciences
  • Individuals who develop and implement AI systems
• Risks in the use of AI: AI systems are implemented in vast, complex environments, the data used and AI will change over time, the model may need to be changed, upgraded, or both, to reflect what is being done with the new data in the environment
• Five dimensions to classify AI systems

  • People and planet
  • Economic context
  • Data and input
  • AI model
  • Tasks and output
• People and planet


  Identifies individuals and groups that might be affected by the AI system. For example, human rights, the environment and society. Privacy comes into play here.
• Economic context


  The AI system is looked at according to the economic and sectoral environment in which it operates.
• Data and input


  What type of data was used in the model and any expert input used.
• AI model


  Discusses the technical type; how the model is built and used.
• Tasks and output


  Tasks that AI systems perform, its outputs and resulting actions from those outputs.
• Categories of AI


  • Recognition
  • Detection
  • Personalization
  • Interaction support
  • Goal-driven optimization
  • Recommendation
• Recognition


  • Facial recognition, retailer product matches, manufacturing machine learning to detect defects, plagiarism detectors
• Detection


  • Credit card transaction fraud detection, events and sports video review, cyber events and systems management
• Personalization


  • Unique online customer profiles, virtual assistants or chatbots
• Goal-driven optimization


  • Supply chain optimization, optimizing driving routes and idle time for vehicles
• Recommendation


  • Product recommendations or viewing recommendations for customers, decision support systems
• Recognizing that terminology can vary will allow you to align how specific terms are used and applied in your organization to ensure that your stakeholders are able to discuss AI use and risk management processes with clarity. Knowing how AI functions and when it is appropriate to use it as well as understanding the types of AI and the ways it can be used, will help determine whether the benefits to your organization of using AI are outweighed by the risks.
• Types of artificial intelligence

  • Artificial Narrow Intelligence (ANI)
  • Artificial General Intelligence (AGI)
  • Artificial Superintelligence (ASI)

  
• Artificial Narrow Intelligence (ANI)


  Designed to perform a single or a narrow set of related tasks at a high level of proficiency
• Artificial General Intelligence (AGI)


  More advanced in scope than ANI, capable of performing a broader set of tasks
• Artificial Superintelligence (ASI)


  AI systems with intellectual powers beyond those of humans across a comprehensive range of categories and fields of endeavor
• Machine learning models


  • Supervised learning
  • Unsupervised learning
  • Reinforcement learning
• Supervised learning models


  Learn from a pre-labeled and classified data set
• Unsupervised learning models


  Do not rely on labeled datasets, designed to identify differences, similarities and other patterns without human supervision
• Reinforcement learning models


  Use a reward and punishment matrix to determine a correct or optimal outcome, rely on trial and error to determine what to do or what not to do