|
on Artificial Intelligence |
| By: | Otis, Nicholas G.; Cranney, Katelyn; Delecourt, Solene; Koning, Rembrand (Harvard Business School) |
| Abstract: | Generative AI has the potential to transform productivity and reduce inequality, but only if used broadly. In this paper, we show that recently identified gender gaps in AI use are nearly universal. Synthesizing evidence from 16 studies that surveyed 100, 000 individuals across 26 countries, along with new data on the gender of AI platform users, we show that the AI gender gap is present in nearly all regions, sectors, and occupations. Using data from two studies that offered participants the chance to use AI tools, we then show that even when the opportunity for men and women to access AI is equalized, women are still less likely to use AI. Our findings underscore the critical need for targeted interventions that go beyond access to address the structural and behavioral barriers that have resulted in a global gender gap in AI use. |
| Date: | 2024–10–14 |
| URL: | https://d.repec.org/n?u=RePEc:osf:osfxxx:h6a7c |
| By: | Ozge Demirci; Jonas Hannane; Xinrong Zhu |
| Abstract: | This paper studies the impact of Generative AI technologies on the demand for online freelancers using a large dataset from a leading global freelancing platform. We identify the types of jobs that are more affected by Generative AI and quantify the magnitude of the heterogeneous impact. Our findings indicate a 21% decrease in the number of job posts for automation-prone jobs related to writing and coding, compared to jobs requiring manual-intensive skills, within eight months after the introduction of ChatGPT. We show that the reduction in the number of job posts increases competition among freelancers while the remaining automation-prone jobs are of greater complexity and offer higher pay. We also find that the introduction of Image-generating AI technologies led to a 17% decrease in the number of job posts related to image creation. We use Google Trends to show that the more pronounced decline in the demand for freelancers within automation-prone jobs correlates with their higher public awareness of ChatGPT’s substitutability. |
| Keywords: | generative AI, large language models, ChatGPT, digital freelancing platforms |
| JEL: | O33 E24 J21 J24 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_11276 |
| By: | Antonio Minniti (Department of Economics, University of Bologna); Klaus Prettner (Department of Economics, Vienna University of Economics and Business); Francesco Venturini (Department of Economics, University of Urbino) |
| Abstract: | We study how the development of Artificial Intelligence (AI) influences the distribution of income between capital and labor and how this, in turn, exacerbates geographic income inequality. To investigate this issue, we first build a theoretical framework and then analyze data from European regions dating back to 2000. We find that for every doubling of regional AI innovation, there is a 0.7% to 1.6% decline in the labor share, which may have decreased by between 0.20 and 0.46 percentage points from a mean of 52% due solely to AI. This new technology is particularly detrimental to high-skill and medium-skill labor. The impact on income distribution is driven by worsening wage and employment conditions for high-skill labor, and by wage compression for medium- and low-skill labor. The effect of AI is not driven by other factors affecting regional development in Europe, nor by the concentration process in the AI market. |
| Keywords: | Artificial Intelligence, patenting, labor share, European regions |
| JEL: | O31 O32 O34 |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:wiw:wiwwuw:wuwp369 |
| By: | Erdem Dogukan Yilmaz; Christian Peukert |
| Abstract: | We examine how the adoption of digital automation technology affects labor demand, operations and profitability in the context of the logistics industry. Our data covers 9, 300 digital automation projects in a multinational company involving service robots and machine learning-based software from 2019 to 2021, alongside fine-grained labor and operations data. To identify causal effects, we leverage exogenous variation from supply-chain disruptions and travel restrictions during COVID-19 and an import ban on information and communication technologies imposed by the Trump administration. We find that total labor cost increased after the adoption of digital automation technology, attributable to increased labor demand and more reliance on temporary workers. However, managerial hours declined, possibly due to increased efficiency. Furthermore, digital automation technology increased revenue and profit through a reduction in operational cost, improved utilization of warehouse space, and higher profit margins. However, the effects of digital automation technology are not homogeneous. We highlight substantial complementarities between hardware and software technologies. Management units that only use software technology experience only half the increase in revenue and profit. |
| Keywords: | digital automation technology, robots, artificial intelligence, future of work |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_11321 |
| By: | Anthony R. Harding; Juan Moreno-Cruz |
| Abstract: | With the rapid expansion of Artificial Intelligence, there are expectations for a proportional expansion of economic activity due to increased productivity, and with it energy consumption and its associated environmental consequences like carbon dioxide emissions. Here, we combine data on economic activity, with early estimates of likely adoption of AI across occupations and industries, to estimate the increase in energy use and carbon dioxide emissions at the industry level and in aggregate for the US economy. At the industry level, energy use can increase between 0 and 12 PJ per year, while emissions increase between 47 tCO2 and 272 ktCO2. Aggregating across industries in the US economy, this totals an increase in energy consumption of 28 PJ per year, or around 0.03% of energy use per year in the US. We find this translates to an increase in carbon dioxide emissions of 896 ktCO2 per year, or around 0.02% of the CO2 emissions per year in the US. |
| Keywords: | artificial intelligence, energy, climate change |
| JEL: | O44 Q43 Q54 Q55 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_11360 |
| By: | Hunt, Jennifer (Rutgers University); Cockburn, Iain (Boston University); Bessen, James (Boston University) |
| Abstract: | Using our own data on Artificial Intelligence publications merged with Burning Glass vacancy data for 2007-2019, we investigate whether online vacancies for jobs requiring AI skills grow more slowly in U.S. locations farther from pre-2007 AI innovation hotspots. We find that a commuting zone which is an additional 200km (125 miles) from the closest AI hotspot has 17% lower growth in AI jobs' share of vacancies. This is driven by distance from AI papers rather than AI patents. Distance reduces growth in AI research jobs as well as in jobs adapting AI to new industries, as evidenced by strong effects for computer and mathematical researchers, developers of software applications, and the finance and insurance industry. 20% of the effect is explained by the presence of state borders between some commuting zones and their closest hotspot. This could reflect state borders impeding migration and thus flows of tacit knowledge. Distance does not capture difficulty of in-person or remote collaboration nor knowledge and personnel flows within multi-establishment firms hiring in computer occupations. |
| Keywords: | Artificial Intelligence, technology adoption and diffusion |
| JEL: | O33 R12 |
| Date: | 2024–09 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp17325 |
| By: | Prithwiraj Choudhury; Bart S. Vanneste; Amirhossein Zohrehvand |
| Abstract: | Can generative artificial intelligence (AI) transform the role of the CEO by effectively automating CEO communication? This study investigates whether AI can mimic a human CEO and whether employees’ perception of the communication’s source matter. In a field experiment with a firm, we extend the idea of a Turing test (i.e., a computer mimicking a person), to the idea of generative AI mimicking a specific person, namely the CEO. We call this the “Wade test” and assess if employees can distinguish between communication from their CEO and communication generated by an AI trained on the CEO’s prior communications. We find that AI responses are correctly identified 59% of the time, somewhat better than random chance. When employees believe a response is AI generated, regardless of its actual source, they perceive it as less helpful. To assess causal mechanisms, a second study with a general audience, using public statements from CEOs and from an AI intended to mimic those CEOs, finds that AI-labeled responses (irrespective of their actual source) are rated as less helpful. These findings highlight that, when using generative AI in CEO communication, people may inaccurately identify the source of communication and exhibit aversion towards communication they identify as being AI generated. |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_11316 |
| By: | Andrea L. Eisfeldt; Gregor Schubert |
| Abstract: | We provide evidence that the development and adoption of Generative AI is driving a significant technological shift for firms and for financial research. We review the literature on the impact of ChatGPT on firm value and provide directions for future research investigating the impact of this major technology shock. Finally, we review and describe innovations in research methods linked to improvements in AI tools, along with their applications. We offer a practical introduction to available tools and advice for researchers interested in using these tools. |
| JEL: | G0 |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:33076 |
| By: | Kelvin J. L. Koa; Yunshan Ma; Ritchie Ng; Huanhuan Zheng; Tat-Seng Chua |
| Abstract: | Stock portfolios are often exposed to rare consequential events (e.g., 2007 global financial crisis, 2020 COVID-19 stock market crash), as they do not have enough historical information to learn from. Large Language Models (LLMs) now present a possible tool to tackle this problem, as they can generalize across their large corpus of training data and perform zero-shot reasoning on new events, allowing them to detect possible portfolio crash events without requiring specific training data. However, detecting portfolio crashes is a complex problem that requires more than basic reasoning abilities. Investors need to dynamically process the impact of each new information found in the news articles, analyze the the relational network of impacts across news events and portfolio stocks, as well as understand the temporal context between impacts across time-steps, in order to obtain the overall aggregated effect on the target portfolio. In this work, we propose an algorithmic framework named Temporal Relational Reasoning (TRR). It seeks to emulate the spectrum of human cognitive capabilities used for complex problem-solving, which include brainstorming, memory, attention and reasoning. Through extensive experiments, we show that TRR is able to outperform state-of-the-art solutions on detecting stock portfolio crashes, and demonstrate how each of the proposed components help to contribute to its performance through an ablation study. Additionally, we further explore the possible applications of TRR by extending it to other related complex problems, such as the detection of possible global crisis events in Macroeconomics. |
| Date: | 2024–10 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2410.17266 |
| By: | Xavier Gabaix; Ralph S J Koijen; Robert Richmond; Motohiro Yogo |
| Abstract: | Asset demand systems specify the demand of investors for financial assets and the supply of securities by firms. We discuss how realistic models of the asset demand system are essential to assess ex post, and predict ex ante, how central bank policy interventions impact asset prices, the distribution of wealth across households and institutions, and financial stability. Due to the improved availability of big holdings data and advances in modelling techniques, estimating asset demand systems is now a practical reality. We show how demand systems provide improved information for policy decisions (eg in the context of financial contagion, convenience yield or the strength of the dollar) or to design optimal policies (eg in the context of quantitative easing or designing climate stress tests). We discuss how recent AI methods can be used to improve models of the asset demand system by better measuring asset and investor similarity through so-called embeddings. These embeddings can for instance be used for policymaking by central banks to understand the rebalancing channel of asset purchase programs and to measure crowded trades. |
| Keywords: | asset prices, central bank policies, artificial intelligence, embeddings |
| JEL: | C5 G11 G12 |
| URL: | https://d.repec.org/n?u=RePEc:bis:biswps:1222 |
| By: | Yikai Zhao; Jun Nagayasu; Xinyi Geng |
| Abstract: | This study examines the impact of climate policy uncertainty (CPU) on credit spreads using data from corporate bonds listed on the Chinese exchange market between 2008 and 2022. We innovatively apply large language models (LLMs) to construct a firm-level CPU index based on disclosure texts and validateits effectiveness. We find that a CPU rise widens a firm’s credit spreads by exacerbating financial distress. Although disclosing environmental, social, and governance (ESG) information moderate CPU’s effect on credit spreads, controversies in ESG ratings amplify it. Finally, heterogeneity analyses reveal that CPU’s effect on wideningbond spreads is more pronounced for traditional bonds, short- to medium-term bonds, nonstate-owned enterprises, and issuing firms with dispersed supply chains. |
| Date: | 2024–11 |
| URL: | https://d.repec.org/n?u=RePEc:toh:dssraa:143 |