1) Technology often automates specific tasks within jobs rather than completely replacing jobs, leading to partial automation and changes in the composition of jobs.
2) While automation may reduce some jobs, it can also create new jobs in related industries to meet increased demand, resulting in job churn rather than widespread unemployment.
3) When machines and humans complement each other's skills, it can lead to occupational substitution as some jobs are replaced while new occupations emerge that combine work with machines. This has contributed to job polarization.
2. Thinking about future tech…
• This time IS different
• Identify what is different
• Understand what is similar
• Propose:
• Look at past
• Recent
• History
• Speculate…
5. Tasks v. Jobs
• Automation has largely automated specific tasks
• Only rarely completely automating jobs
• Jobs involve many tasks requiring diverse skills
8. Machine Learning: Better than humans?
• Reading lips
• Playing chess, Go, Jeopardy
• Reading X-rays
9. AI & Complete Automation
• Machine Learning = “Idiot Savant”
• Individual tasks
• Large, labelled data
• No model of the world, common sense, meaning
• Basic method around since 1980s
• Big Data, Big Hardware
• “General Artificial Intelligence”
• AI as marketing
10. Accountants &
auditors Loan Officers
Paralegals &
legal assistants
First computer
automation
1954 1956; AI 1987 1963
What is being
developed?
Scan, categorize, &
enter data from paper;
retrieve contract terms
Automatic loan
processing; fraud
detection
Legal text retrieval;
e-discovery
Task ML can't do
Evaluate and advise re
organizational controls
Monitor business
performance evaluating
non-standardized
information
Evaluate retrieved
documents for legal
relevance to case
Job growth, 2000-16 10% -- 13%
Predicted, 2016-26
(BLS)
10% 11% 15%
Examples: Partial automation
39. Closed knowledge
• Proprietary systems
• Walmart logistics
• Large bank credit card systems
• Boeing/Airbus design systems
• Why hard to imitate?
• Economies of scale, network effects
• Proprietary knowledge
• Hiring
• Training
• Learning on the job
40. Closed Knowledge & Wages
• Weaker labor markets for technology-specific skills
• Limited alternative employment
• Weaker incentives to invest in skills
• Lower wage growth
41. Knowledge & wages historically
• Automation of tasks
• Increases value of skill on remaining tasks
• Greater demand for skill
• Demand higher wages
• BUT ONLY with robust labor markets