|
on Artificial Intelligence |
| By: | Joshua S. Gans |
| Abstract: | This paper examines recent proposals and research suggesting that AI adoption should be delayed until its potential harms are properly understood. It is shown that conclusions regarding the social optimality of delayed AI adoption are sensitive to assumptions regarding the process by which regulators learn about the salience of particular harms. When such learning is by doing -- based on the real-world adoption of AI -- this generally favours acceleration of AI adoption to surface and react to potential harms more quickly. This case is strengthened when AI adoption is potentially reversible. The paper examines how different conclusions regarding the optimality of accelerated or delayed AI adoption influence and are influenced by other policies that may moderate AI harm. |
| JEL: | L51 O33 |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:32105&r=ain |
| By: | Paolo Carioli; Dirk Czarnitzki; Gastón P Fernández Barros |
| Abstract: | Artificial Intelligence (AI) is considered to be the next general-purpose technology, with the potential of performing tasks commonly requiring human capabilities. While it is commonly feared that AI replaces labor and disrupts jobs, we instead investigate the potential of AI for overcoming increasingly alarming skills shortages in firms. We exploit unique German survey data from the Mannheim Innovation Panel on both the adoption of AI and the extent to which firms experience scarcity of skills. We measure skills shortage by the number of job vacancies that could not be filled as planned by firms, distinguishing among different types of skills. To account for the potential endogeneity of skills shortage, we also implement instrumental variable estimators. Overall, we find a positive and significant effect of skills shortage on AI adoption, the breadth of AI methods, and the breadth of areas of application of AI. In addition, we find evidence that scarcity of labor with academic education relates to firms exploring and adopting AI. |
| Keywords: | Artificial Intelligence, CIS data, skills shortage |
| Date: | 2024–02–08 |
| URL: | http://d.repec.org/n?u=RePEc:ete:ecoomp:735893&r=ain |
| By: | David Autor |
| Abstract: | While the utopian vision of the current Information Age was that computerization would flatten economic hierarchies by democratizing information, the opposite has occurred. Information, it turns out, is merely an input into a more consequential economic function, decision-making, which is the province of elite experts. The unique opportunity that AI offers to the labor market is to extend the relevance, reach, and value of human expertise. Because of AI’s capacity to weave information and rules with acquired experience to support decision-making, it can be applied to enable a larger set of workers possessing complementary knowledge to perform some of the higher-stakes decision-making tasks that are currently arrogated to elite experts, e.g., medical care to doctors, document production to lawyers, software coding to computer engineers, and undergraduate education to professors. My thesis is not a forecast but an argument about what is possible: AI, if used well, can assist with restoring the middle-skill, middle-class heart of the US labor market that has been hollowed out by automation and globalization. |
| JEL: | J01 J2 N30 O14 |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:32140&r=ain |
| By: | Joshua S. Gans |
| Abstract: | New generative artificial intelligence (AI) models, including large language models and image generators, have created new challenges for copyright policy as such models may be trained on data that includes copy-protected content. This paper examines this issue from an economics perspective and analyses how different copyright regimes for generative AI will impact the quality of content generated as well as the quality of AI training. A key factor is whether generative AI models are small (with content providers capable of negotiations with AI providers) or large (where negotiations are prohibitive). For small AI models, it is found that giving original content providers copyright protection leads to superior social welfare outcomes compared to having no copyright protection. For large AI models, this comparison is ambiguous and depends on the level of potential harm to original content providers and the importance of content for AI training quality. However, it is demonstrated that an ex-post `fair use' type mechanism can lead to higher expected social welfare than traditional copyright regimes. |
| JEL: | K20 O34 |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:32106&r=ain |
| By: | Jean Lee; Nicholas Stevens; Soyeon Caren Han; Minseok Song |
| Abstract: | Large Language Models (LLMs) have shown remarkable capabilities across a wide variety of Natural Language Processing (NLP) tasks and have attracted attention from multiple domains, including financial services. Despite the extensive research into general-domain LLMs, and their immense potential in finance, Financial LLM (FinLLM) research remains limited. This survey provides a comprehensive overview of FinLLMs, including their history, techniques, performance, and opportunities and challenges. Firstly, we present a chronological overview of general-domain Pre-trained Language Models (PLMs) through to current FinLLMs, including the GPT-series, selected open-source LLMs, and financial LMs. Secondly, we compare five techniques used across financial PLMs and FinLLMs, including training methods, training data, and fine-tuning methods. Thirdly, we summarize the performance evaluations of six benchmark tasks and datasets. In addition, we provide eight advanced financial NLP tasks and datasets for developing more sophisticated FinLLMs. Finally, we discuss the opportunities and the challenges facing FinLLMs, such as hallucination, privacy, and efficiency. To support AI research in finance, we compile a collection of accessible datasets and evaluation benchmarks on GitHub. |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2402.02315&r=ain |
| By: | Kelvin J. L. Koa; Yunshan Ma; Ritchie Ng; Tat-Seng Chua |
| Abstract: | Explaining stock predictions is generally a difficult task for traditional non-generative deep learning models, where explanations are limited to visualizing the attention weights on important texts. Today, Large Language Models (LLMs) present a solution to this problem, given their known capabilities to generate human-readable explanations for their decision-making process. However, the task of stock prediction remains challenging for LLMs, as it requires the ability to weigh the varying impacts of chaotic social texts on stock prices. The problem gets progressively harder with the introduction of the explanation component, which requires LLMs to explain verbally why certain factors are more important than the others. On the other hand, to fine-tune LLMs for such a task, one would need expert-annotated samples of explanation for every stock movement in the training set, which is expensive and impractical to scale. To tackle these issues, we propose our Summarize-Explain-Predict (SEP) framework, which utilizes a self-reflective agent and Proximal Policy Optimization (PPO) to let a LLM teach itself how to generate explainable stock predictions in a fully autonomous manner. The reflective agent learns how to explain past stock movements through self-reasoning, while the PPO trainer trains the model to generate the most likely explanations from input texts. The training samples for the PPO trainer are also the responses generated during the reflective process, which eliminates the need for human annotators. Using our SEP framework, we fine-tune a LLM that can outperform both traditional deep-learning and LLM methods in prediction accuracy and Matthews correlation coefficient for the stock classification task. To justify the generalization capability of our framework, we further test it on the portfolio construction task, and demonstrate its effectiveness through various portfolio metrics. |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2402.03659&r=ain |
| By: | Teodoro Baldazzi (Università Roma Tre); Luigi Bellomarini (Bank of Italy); Stefano Ceri (Politecnico di Milano); Andrea Colombo (Politecnico di Milano); Andrea Gentili (Bank of Italy); Emanuel Sallinger (TU Wien; University of Oxford) |
| Abstract: | Large Language Models (LLMs) usually undergo a pre-training process on extensive collections of generic textual data, which are often publicly accessible. Pre-training enables LLMs to grasp language grammar, understand context, and convey a sense of common knowledge. Pre-training can be likened to machine learning training: the LLM is trained to predict the next basic text unit (e.g., a word or a sequence of words) based on the sequence of previously observed units. However, despite the impressive generalization and human-like interaction capabilities shown in Natural Language Processing (NLP) tasks, pre-trained LLMs exhibit significant limitations and provide poor accuracy when applied in specialized domains. Their main limitation stems from the fact that data used in generic pre-training often lacks knowledge related to the specific domain. To address these limitations, fine-tuning techniques are often employed to refine pre-trained models using domain-specific data. Factual information is extracted from company databases to create text collections for fine-tuning purposes. However, even in this case, results tend to be unsatisfactory in complex domains, such as financial markets and finance in general. Examining the issue from a different perspective, the Knowledge Representation and Reasoning (KRR) community has focused on producing formalisms, methods, and systems for representing complex Enterprise Knowledge. In particular, Enterprise Knowledge Graphs (EKGs) can leverage a combination of factual information in databases and business knowledge specified in a compact and formal fashion. EKGs serve the purpose of answering specific domain queries through established techniques such as ontological reasoning. Domain knowledge is represented in symbolic forms, e.g., logic-based languages, and used to draw consequential conclusions from the available data. However, while EKGs are applied successfully in many financial scenarios, they lack flexibility, common sense and linguistic orientation, essential for NLP. This paper proposes an approach aimed at enhancing the utility of LLMs for specific applications, such as those related to financial markets. The approach involves guiding the fine-tuning process of LLMs through ontological reasoning on EKGs. In particular, we exploit the Vadalog system and its language, a state-of-the-art automated reasoning framework, to synthesize an extensive fine- tuning corpus from a logical formalization of domain knowledge in an EKG. Our contribution consists of a technique called verbalization, which transforms the set of inferences determined by ontological reasoning into a corpus for fine-tuning. We present a complete software architecture that applies verbalization to four NLP tasks: question answering, i.e., providing accurate responses in a specific domain in good prose; explanation, i.e., systematically justifying the conclusions drawn; translation, i.e., converting domain specifications into logical formalization; and description, i.e., explaining formal specifications in prose. We apply the approach and our architecture in the context of financial markets, presenting a proof of concept that highlights their advantages. |
| Keywords: | Ontological reasoning, Large language models, Knowledge graphs |
| Date: | 2024–01 |
| URL: | http://d.repec.org/n?u=RePEc:bdi:wpmisp:mip_044_24&r=ain |
| By: | Saizhuo Wang; Hang Yuan; Lionel M. Ni; Jian Guo |
| Abstract: | Autonomous agents based on Large Language Models (LLMs) that devise plans and tackle real-world challenges have gained prominence.However, tailoring these agents for specialized domains like quantitative investment remains a formidable task. The core challenge involves efficiently building and integrating a domain-specific knowledge base for the agent's learning process. This paper introduces a principled framework to address this challenge, comprising a two-layer loop.In the inner loop, the agent refines its responses by drawing from its knowledge base, while in the outer loop, these responses are tested in real-world scenarios to automatically enhance the knowledge base with new insights.We demonstrate that our approach enables the agent to progressively approximate optimal behavior with provable efficiency.Furthermore, we instantiate this framework through an autonomous agent for mining trading signals named QuantAgent. Empirical results showcase QuantAgent's capability in uncovering viable financial signals and enhancing the accuracy of financial forecasts. |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2402.03755&r=ain |
| By: | ZHU Chen; MOTOHASHI Kazuyuki |
| Abstract: | Startups have emerged as pivotal innovators in the commercialization of AI technology. Nonetheless, these nascent enterprises often require substantial capital infusion to realize the economic returns from their innovations. This study examines the role of prototypes in facilitating their fundraising process. We utilized historical web content to identify the presence of prototypes and employed web traffic data to monitor their customer growth. Our findings indicate that prototyping positively affects the potential customer attraction process, signaling the feasibility and profitability of their business hypotheses to potential investors. In addition, as a technologically intensive industry, most AI startups begin with a technology-centric approach. While a technology-led starting point underscores competitiveness, it also inherently introduces uncertainty. We offer quantitative evidence demonstrating how prototyping acts as a moderating factor, reducing the impact of such uncertainty by expediting investor decision-making. |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:eti:dpaper:24021&r=ain |
| By: | Xiaorui Zuo; Yao-Tsung Chen; Wolfgang Karl H\"ardle |
| Abstract: | In the burgeoning realm of cryptocurrency, social media platforms like Twitter have become pivotal in influencing market trends and investor sentiments. In our study, we leverage GPT-4 and a fine-tuned transformer-based BERT model for a multimodal sentiment analysis, focusing on the impact of emoji sentiment on cryptocurrency markets. By translating emojis into quantifiable sentiment data, we correlate these insights with key market indicators like BTC Price and the VCRIX index. This approach may be fed into the development of trading strategies aimed at utilizing social media elements to identify and forecast market trends. Crucially, our findings suggest that strategies based on emoji sentiment can facilitate the avoidance of significant market downturns and contribute to the stabilization of returns. This research underscores the practical benefits of integrating advanced AI-driven analyses into financial strategies, offering a nuanced perspective on the interplay between digital communication and market dynamics in an academic context. |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2402.10481&r=ain |