Artificial Intelligence and Machine Learning

Question 1

Artificial intelligence goes hand in hand with big data?

Answer 1

TRUE

Question 2

What are categories of AI?

Answer 2

1) Machine learning- Use big data to learn rules and categories to enable prediction and classification. Common for accounting applications- classifying journal entries.
2) Robotics- Directed observation and activity (physical tasks)- welding, controlling production, manufacturing and distributing processes.
3) Intelligent Agents- Computer “agents” that perform tasks. Analysis of market trends (in purchasing airline tickets). Interact with humans (Siri, Cortana)
4) Expert Systems- build and apply expertise in a woman (configuration Dell Systems, medical diagnosis). May include learning or intelligent agent subsystems.
5) AI Tasks and Intelligence- Data harvesting and cleaning. Prepping and cleaning client data (loan documents). Analyzing numbers suggest what to watch next based on quantitative data about your ratings and viewing habits. Analyzing words, images, and sounds.

Question 3

AI accounting applications- reported big 4 initiatives?

Answer 3

• KPMG- Partner with IBM Watson initiative (expert system and natural language processing system). Use Watson to identify F/S risks.
• Deloitte- Partnering with number of vendors to automate audit tasks and processes including document review, confirmations, inventory counts, disclosures, predictive risk analytics, and client request lists (Kira Systems)
• PwC- Halo system analyze accounting journal entries for risky transactions and identify inefficient processes. Automates coding and classification of journal entries
• EY- Focuses on integrating Big Data and advanced analytics into audit practice.

Question 4

What are AI benefits?

Answer 4

Speed, accuracy, and cost

• Ability to scale and speed applications, reduced cots, new applications
• Ability to obtain, clean, and analyze large data in real time
• Apply robotics, pattens recognition

Question 5

What are AI risks?

Answer 5

AI systems often include the biases of their developers

• Data biases- harvesting and creating data sets that omit relevant variables and considerations
• Prediction biases- systems that include biased data will, obviously, generate biased predictions
• Learned biases- smart machines will learn and integrate the biases of those who train them
• Lost jobs and disappearing employment
• Legal and ethical issues

Long term:
- The end of humanity? Machine overloads?
-

Question 6

What are the implications of AI on accounting?

Answer 6

Evolution of accounting jobs include

• Partnering with AI to monitor and improve results
• Oversight of AI in accounting and auditing tasks
• When and how tasks should be automated vs not
• Developing new and refining existing AI applications
• Executing human only tasks