Embedded Files
Notice
Notice
PEW in 2025 will be held on the fourth Saturday of each month. The workshop starts at 3:00 p.m.
Each workshop basically consists of two sessions: a 1-hour presentation and a 1.5-hour presentation by an invited speaker.
Next Workshop
Next Workshop
November 22 (Sat.) 2025
Venue: 709 in Building 3, Waseda Campus, Waseda University
1st Session
Speaker: Kazuki Nishikawa, Graduate School of Economics, The University of Osaka
Title: Automation under relational contracts
Abstract:
I examine firms’ incentives to invest in automation under relational contracts. In these contracts, the lack of formal enforcement mechanisms can lead to worker shirking, thereby strengthening the incentive to automate. At the same time, automation may weaken the firm–worker relationship, creating a countervailing incentive to limit automation in order to preserve its value. The interplay of these opposing forces determines the relative level of automation investment under relational versus formal contracts. From a policy perspective, the results indicate that strengthening workers’ rights does not necessarily raise workers’ surplus.
Language: Japanese
Time: 15:00 - 16:00
2nd Session
Speaker: Xuanli Zhu, Keio University
Title: Where is the Bottleneck? The Productivity Paradox of Generative AI
Abstract:
Recent advances in generative artificial intelligence (GenAI) tools raise concerns about a new productivity paradox of general technology. However, there is little systematic evidence on how large the gap is between GenAI’s theoretical efficiency and actual usage by human workers, and what the key bottlenecks are underlying this productivity gap. By systematically prompting frontierreasoning AI models to predict time savings across tasks and occupations, a new method that has gained rapid popularity in recent AI literature, we find GenAI predicts a positive productivity effect for most occupations, with an average ranging from 11% to 34% depending on prompt design and aggregation methods. Compared to real-world human surveys, this suggests that a large productivity gap exists due to both limited adoption (extensive margin) and inefficient usage (intensive margin). We then conduct counterfactual experiments by varying prompt design to identify various potential bottlenecks. Results reveal that human overhead costs of managing AI and a lack of specialized tools integrated into workflows serve as important constraints to productivity gains, whereas recent AI capability improvements yield minimal additional predicted time savings.
Language: English
Time: 16:15-17:45
How to attend?
How to attend?
Sign up to our mailing list to receive information about our workshop by sending an email to: eoe-office[@]list.waseda.jp (Replace [@] with @.)
In-person venues or/and Zoom meeting URL will be announced via the mailing list.
Upcoming Workshops
Upcoming Workshops
December 13 (Sat.) 2025
Venue: 203 in Building 7, Waseda Campus, Waseda University
1st Session
Speaker:TBA
Title: TBA
Abstract:
TBA
Language: TBA
Time: TBA
2nd Session
Speaker: TBA
Title: TBA
Abstract:
TBA
Language: TBA
Time: TBA
Page updated
Google Sites
Report abuse