File 22: Artificial intelligence

Question 1

What are the two ways that programming led to computers defeatings world class players of the game? (2) According to Frey

Answer 1

1. Deep Blue of IBM = chess game: programmers look at games played by grand masters and programmed the decisions the computer should make
2. Alpha Go: provided with lots of records of game played=> instructed the computer to play against itself thus learning from experience = machine learning

Question 2

What does machine learning depend on? (4) What is each’s potential employment effect

Answer 2

data sets:image recongnition (security camera) - security agents
, translation - translators
, voice recognition - call centers
vehicle performance -autonomous driveing replacing trucks

Question 3

What are Frey’s 3 ideas to interpret the likely effects of machine learning?

Answer 3

1. POLANYI’S PARADOX: human’s task draw on a very large reservoir of tacit knowledge rather than/ or along with formal knowledge
2. MORAVEC’S PARADOX: computer/automated equipment have a hard time doing tasks that are easy for humans (folding a towel, playing soccer, setting a chess board)
3. AMAR’S LAW: productivity gains from new technology require significant complementary investments: tractors became available to replace horses in teh early 20th century but their use diffused slowly: bc their effective use required substantial expenditure on training operators and ways to broaden use of tractor (first used for tillage then for sowing, harvesting => dev of addition fittings for the tractors)

Question 4

How do frey’s 3 ideas on the effects of machine learning affect the adoption of AI and its employment effects?

Answer 4

Polanyi’s paradox limits AI because task require tacit knowledge
moravec’s paradox is overcome by machine learning
Amar’s law says that productivity gains from machine learning will be slowed by the need for complementary investment => autonomous vehicledepend on investments in signalling along roadways

Question 5

What kind of jobs that are more likely vulnerable to innovations assoiciated with machine learning? give Frey’s examples (4)

Answer 5

1. require less tacit knowledge
2. will involve task that can be simplified
3. require smaller complemetary investments
   ex: food preparation and serving, production (robots in auto plants), contrusction thru pre fabrication, farming/fishing and forestry