|
on Artificial Intelligence |
| By: | Nicole Immorlica; Brendan Lucier; Aleksandrs Slivkins |
| Abstract: | Traditionally, AI has been modeled within economics as a technology that impacts payoffs by reducing costs or refining information for human agents. Our position is that, in light of recent advances in generative AI, it is increasingly useful to model AI itself as an economic agent. In our framework, each user is augmented with an AI agent and can consult the AI prior to taking actions in a game. The AI agent and the user have potentially different information and preferences over the communication, which can result in equilibria that are qualitatively different than in settings without AI. |
| Date: | 2024–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2406.00477&r= |
| By: | Bernhard Kasberger; Simon Martin; Hans-Theo Normann; Tobias Werner |
| Abstract: | Algorithms play an increasingly important role in economic situations. These situations are often strategic, where the artificial intelligence may or may not be cooperative. We study the deter-minants and forms of algorithmic cooperation in the infinitely repeated prisoner’s dilemma. We run a sequence of computational experiments, accompanied by additional repeated prisoner’s dilemma games played by humans in the lab. We find that the same factors that increase human cooperation largely also determine the cooperation rates of algorithms. However, algorithms tend to play different strategies than humans. Algorithms cooperate less than humans when cooperation is very risky or not incentive-compatible. |
| Keywords: | artificial intelligence, cooperation, large language models, Q-learning, repeated prisoner’s dilemma |
| JEL: | C72 C73 C92 D83 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_11124&r= |
| By: | Berger, Benedikt; Adam, Martin; Rühr, Alexander; Benlian, Alexander |
| Abstract: | Owing to advancements in artificial intelligence (AI) and specifically in machine learning, information technology (IT) systems can support humans in an increasing number of tasks. Yet, previous research indicates that people often prefer human support to support by an IT system, even if the latter provides superior performance – a phenomenon called algorithm aversion. A possible cause of algorithm aversion put forward in literature is that users lose trust in IT systems they become familiar with and perceive to err, for example, making forecasts that turn out to deviate from the actual value. Therefore, this paper evaluates the effectiveness of demonstrating an AI-based system’s ability to learn as a potential countermeasure against algorithm aversion in an incentive-compatible online experiment. The experiment reveals how the nature of an erring advisor (i.e., human vs. algorithmic), its familiarity to the user (i.e., unfamiliar vs. familiar), and its ability to learn (i.e., non-learning vs. learning) influence a decision maker’s reliance on the advisor’s judgement for an objective and non-personal decision task. The results reveal no difference in the reliance on unfamiliar human and algorithmic advisors, but differences in the reliance on familiar human and algorithmic advisors that err. Demonstrating an advisor’s ability to learn, however, offsets the effect of familiarity. Therefore, this study contributes to an enhanced understanding of algorithm aversion and is one of the first to examine how users perceive whether an IT system is able to learn. The findings provide theoretical and practical implications for the employment and design of AI-based systems. |
| Date: | 2024–06–18 |
| URL: | https://d.repec.org/n?u=RePEc:dar:wpaper:146095&r= |
| By: | Jingru Jia; Zehua Yuan; Junhao Pan; Paul McNamara; Deming Chen |
| Abstract: | When making decisions under uncertainty, individuals often deviate from rational behavior, which can be evaluated across three dimensions: risk preference, probability weighting, and loss aversion. Given the widespread use of large language models (LLMs) in decision-making processes, it is crucial to assess whether their behavior aligns with human norms and ethical expectations or exhibits potential biases. Several empirical studies have investigated the rationality and social behavior performance of LLMs, yet their internal decision-making tendencies and capabilities remain inadequately understood. This paper proposes a framework, grounded in behavioral economics, to evaluate the decision-making behaviors of LLMs. Through a multiple-choice-list experiment, we estimate the degree of risk preference, probability weighting, and loss aversion in a context-free setting for three commercial LLMs: ChatGPT-4.0-Turbo, Claude-3-Opus, and Gemini-1.0-pro. Our results reveal that LLMs generally exhibit patterns similar to humans, such as risk aversion and loss aversion, with a tendency to overweight small probabilities. However, there are significant variations in the degree to which these behaviors are expressed across different LLMs. We also explore their behavior when embedded with socio-demographic features, uncovering significant disparities. For instance, when modeled with attributes of sexual minority groups or physical disabilities, Claude-3-Opus displays increased risk aversion, leading to more conservative choices. These findings underscore the need for careful consideration of the ethical implications and potential biases in deploying LLMs in decision-making scenarios. Therefore, this study advocates for developing standards and guidelines to ensure that LLMs operate within ethical boundaries while enhancing their utility in complex decision-making environments. |
| Date: | 2024–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2406.05972&r= |
| By: | Can Celebi (University of Mannheim); Stefan Penczynski (School of Economics and Centre for Behavioural and Experimental Social Science, University of East Anglia) |
| Abstract: | In our study, we compare the classification capabilities of GPT-3.5 and GPT-4 with human annotators using text data from economic experiments. We analysed four text corpora, focusing on two domains: promises and strategic reasoning. Starting with prompts close to those given to human annotators, we subsequently explored alternative prompts to investigate the effect of varying classification instructions and degrees of background information on the models’ classification performance. Additionally, we varied the number of examples in a prompt (few-shot vs zero-shot) and the use of the zero-shot “Chain of Thought†prompting technique. Our findings show that GPT-4’s performance is comparable to human annotators, achieving accuracy levels near or over 90% in three tasks, and in the most challenging task of classifying strategic thinking in asymmetric coordination games, it reaches an accuracy level above 70%. |
| Keywords: | Text Classification, GPT, Strategic Thinking, Promises |
| Date: | 2024–06 |
| URL: | https://d.repec.org/n?u=RePEc:uea:wcbess:24-01&r= |
| By: | Simon D Angus (SoDa Laboratories & Dept. of Economics, Monash Business School); Lachlan O'Neill (SoDa Laboratories, Monash Business School) |
| Abstract: | Detecting and quantifying issue framing in textual discourse - the slant or perspective one takes to a given topic (e.g. climate science vs. denialism, misogyny vs. gender equality) - is highly valuable to a range of end-users from social and political scientists to program evaluators and policy analysts. Being able to identify statistically significant shifts, reversals, or changes in issue framing in public discourse would enable the quantitative evaluation of interventions, actors and events that shape discourse. However, issue framing is notoriously challenging for automated natural language processing (NLP) methods since the words and phrases used by either 'side' of an issue are often held in common, with only subtle stylistic flourishes separating their use. Here we develop and rigorously evaluate new detection methods for issue framing and narrative analysis within large text datasets. By introducing a novel application of next-token log probabilities derived from generative large language models (LLMs) we show that issue framing can be reliably and efficiently detected in large corpora with only a few examples of either perspective on a given issue, a method we call 'paired completion'. Through 192 independent experiments over three novel, synthetic datasets, we evaluate paired completion against prompt-based LLM methods and labelled methods using traditional NLP and recent LLM contextual embeddings. We additionally conduct a cost-based analysis to mark out the feasible set of performant methods at production-level scales, and a model bias analysis. Together, our work demonstrates a feasible path to scalable, accurate and low-bias issue-framing in large corpora. |
| Keywords: | slant detection, text-as-data, synthetic data, computational linguistics |
| JEL: | C19 C55 |
| Date: | 2024–06 |
| URL: | https://d.repec.org/n?u=RePEc:ajr:sodwps:2024-02&r= |
| By: | Eiji Yamamura; Fumio Ohtake |
| Abstract: | Using individual-level survey data from 2024, this study investigated how respondent characteristics are associated with a subjective view of generative artificial intelligence (GAI). We asked 14 questions concerning respondents view about GAI, such as general view, faulty GAI, autonomous GEI, GAI replacing humans, and importance of human learning. Regression analysis based on the ordered logit model revealed that: (1) In some cases, the results of smartphone and computer usage times differed. Smartphone usage time was negatively correlated with the importance of human learning, whereas computer usage was not negatively correlated. (2) Managers and ordinary businesspeople have positive views of GAI. However, managers do not show a positive view about GAI being responsible for human decision making. (3) Teachers generally have a negative view about GAI replacing humans and no need of learning. They do not have negative views about GAI producing documents unless GAI is faulty. (4) Medical industry workers positively view GAI if it operates following their direction. However, they do not agree with the view that GAI replaces humans, and that human learning is unnecessary. (5) Females are less likely than men to have a positive view of GAI. In summary, views about GAI vary widely by the individual characteristics and condition of GAI, and by the question set. |
| Date: | 2024–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2406.03823&r= |
| By: | OECD |
| Abstract: | False and misleading content online poses significant risks to the well-being of people and society, but a lack of cross-country comparable evidence persists. This paper contributes to the literature by presenting the OECD Truth Quest Survey methodology and key findings. This survey assesses whether some types of content are more easily distinguishable as false and misleading than others and whether the theme plays any role in its detection. It provides evidence about whether AI-generated content is easier to identify than human-generated content as well as insights into the effects of AI labelling. It further presents information on people’s behaviour as they interact with false and misleading content and their perceptions about their ability to detect it. The cross-country comparable data from the survey will help policy makers better design media literacy strategies, programmes and related policies to address the negative effects of such content. |
| Keywords: | disinformation, false and misleading content, misinformation |
| Date: | 2024–06–28 |
| URL: | https://d.repec.org/n?u=RePEc:oec:stiaab:369-en&r= |
| By: | Kene Boun My; Julien Jacob; Mathieu Lefebvre |
| Abstract: | We propose a new theoretical framework to analyze the incentives provided by different allocations of liability in the case of (semi)autonomous devices which are a source of risk of accident. We consider three key agents, an AI provider (scientist), a producer and a consumer, and look at the effect of different rules of sharing liability on the decision making of each type of agent. In addition we test the theoretical predictions in an original lab experiment. We show that liability on the scientist and the producer is efficient in reducing their misbehaviors. We also find that liability on the consumer increases her incentives to control the risk of an accident (in case of a semi-autonomous device). However, the absence of consumer’s control (full autonomous device) and liability decreases the consumer’s propensity to buy the good. We complete our study by making a social welfare analysis. It highlights the importance of letting the producer liable in order to provide the consumer with confidence in the technology, especially in the case of a full autonomy of the good. |
| Keywords: | AI, Liability Sharing Rules, asymmetric information, lab experiment. |
| JEL: | C91 D82 K13 K32 |
| Date: | 2024 |
| URL: | https://d.repec.org/n?u=RePEc:ulp:sbbeta:2024-24&r= |
| By: | Koch, Michael (Aarhus University); Lodefalk, Magnus (Örebro University School of Business) |
| Abstract: | We use individual survey data providing detailed information on stress, technology adoption, and work, worker, and employer characteristics, in combination with recent measures of AI and robot exposure, to investigate how new technologies affect worker stress. We find a persistent negative relationship, suggesting that AI and robots could reduce the stress level of workers. We furthermore provide evidence on potential mechanisms to explain our findings. Overall, the results provide suggestive evidence of modern technologies changing the way we perform our work in a way that reduces stress and work pressure. |
| Keywords: | Artificial intelligence technologies; Automation; Task content; Skills; Stress |
| JEL: | I31 J24 J28 J44 N34 O33 |
| Date: | 2024–06–14 |
| URL: | https://d.repec.org/n?u=RePEc:hhs:oruesi:2024_005&r= |
| By: | Mauro Cazzaniga; Carlo Pizzinelli; Emma J Rockall; Ms. Marina Mendes Tavares |
| Abstract: | We document historical patterns of workers' transitions across occupations and over the life-cycle for different levels of exposure and complementarity to Artificial Intelligence (AI) in Brazil and the UK. In both countries, college-educated workers frequently move from high-exposure, low-complementarity occupations (those more likely to be negatively affected by AI) to high-exposure, high-complementarity ones (those more likely to be positively affected by AI). This transition is especially common for young college-educated workers and is associated with an increase in average salaries. Young highly educated workers thus represent the demographic group for which AI-driven structural change could most expand opportunities for career progression but also highly disrupt entry into the labor market by removing stepping-stone jobs. These patterns of “upward” labor market transitions for college-educated workers look broadly alike in the UK and Brazil, suggesting that the impact of AI adoption on the highly educated labor force could be similar across advanced economies and emerging markets. Meanwhile, non-college workers in Brazil face markedly higher chances of moving from better-paid high-exposure and low-complementarity occupations to low-exposure ones, suggesting a higher risk of income loss if AI were to reduce labor demand for the former type of jobs. |
| Keywords: | Artificial intelligence; Employment; Occupations; Emerging Markets |
| Date: | 2024–06–07 |
| URL: | https://d.repec.org/n?u=RePEc:imf:imfwpa:2024/116&r= |
| By: | Fossen, Frank M. (University of Nevada, Reno); McLemore, Trevor (University of Nevada, Reno); Sorgner, Alina (John Cabot University) |
| Abstract: | This survey reviews emerging but fast-growing literature on impacts of artificial intelligence (AI) on entrepreneurship, providing a resource for researchers in entrepreneurship and neighboring disciplines. We begin with a review of definitions of AI and show that ambiguity and broadness of definitions adopted in empirical studies may result in obscured evidence on impacts of AI on en-trepreneurship. Against this background, we present and discuss existing theory and evidence on how AI technologies affect entrepreneurial opportunities and decision-making under uncertainty, the adoption of AI technologies by startups, entry barriers, and the performance of entrepreneurial businesses. We add an original empirical analysis of survey data from the German Socio-economic Panel revealing that entrepreneurs, particularly those with employees, are aware of and use AI technologies significantly more frequently than paid employees. Next, we discuss how AI may affect entrepreneurship indirectly through impacting local and sectoral labor markets. The reviewed evidence suggests that AI technologies that are designed to automate jobs are likely to result in a higher level of necessity entrepreneurship in a region, whereas AI technologies that transform jobs without necessarily displacing human workers increase the level of opportunity entrepreneurship. More generally, AI impacts regional entrepreneurship ecosystems (EE) in multiple ways by altering the importance of existing EE elements and processes, creating new ones, and potentially reducing the role of geography for entrepreneurship. Lastly, we address the question of how regulation of AI may affect the entrepreneurship landscape by focusing on the case of the European Union that has pioneered data protection and AI legislation. We conclude our survey by discussing implications for entrepreneurship research and policy. |
| Keywords: | artificial intelligence, machine learning, entrepreneurship, AI startups, digital entrepreneurship, opportunity, innovation, entrepreneurship ecosystem, digital entrepreneurship ecosystem, AI regulation |
| JEL: | J24 L26 O30 |
| Date: | 2024–06 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp17055&r= |
| By: | Tohid Atashbar |
| Abstract: | Learning from the past is critical for shaping the future, especially when it comes to economic policymaking. Building upon the current methods in the application of Reinforcement Learning (RL) to the large language models (LLMs), this paper introduces Reinforcement Learning from Experience Feedback (RLXF), a procedure that tunes LLMs based on lessons from past experiences. RLXF integrates historical experiences into LLM training in two key ways - by training reward models on historical data, and by using that knowledge to fine-tune the LLMs. As a case study, we applied RLXF to tune an LLM using the IMF's MONA database to generate historically-grounded policy suggestions. The results demonstrate RLXF's potential to equip generative AI with a nuanced perspective informed by previous experiences. Overall, it seems RLXF could enable more informed applications of LLMs for economic policy, but this approach is not without the potential risks and limitations of relying heavily on historical data, as it may perpetuate biases and outdated assumptions. |
| Keywords: | LLMs; GAI; RLHF; RLAIF; RLXF |
| Date: | 2024–06–07 |
| URL: | https://d.repec.org/n?u=RePEc:imf:imfwpa:2024/114&r= |
| By: | Fernanda Brollo; Ms. Era Dabla-Norris; Mr. Ruud de Mooij; Mr. Daniel Garcia-Macia; Tibor Hanappi; Ms. Li Liu; Anh D. M. Nguyen |
| Abstract: | Generative artificial intelligence (gen AI) holds immense potential to boost productivity growth and advance public service delivery, but it also raises profound concerns about massive labor disruptions and rising inequality. This note discusses how fiscal policies can be employed to steer the technology and its deployment in ways that serve humanity best while cushioning the negative labor market and distributional effects to broaden the gains. Given the vast uncertainty about the nature, impact, and speed of developments in gen AI, governments should take an agile approach that prepares them for both business as usual and highly disruptive scenarios. |
| Keywords: | Artificial intelligence; AI; technological change; labor market; fiscal policy; social protection systems; tax systems |
| Date: | 2024–06–17 |
| URL: | https://d.repec.org/n?u=RePEc:imf:imfsdn:2024/002&r= |
| By: | Alexander Bakumenko (Clemson University, USA); Kate\v{r}ina Hlav\'a\v{c}kov\'a-Schindler (University of Vienna, Austria); Claudia Plant (University of Vienna, Austria); Nina C. Hubig (Clemson University, USA) |
| Abstract: | Detecting anomalies in general ledger data is of utmost importance to ensure trustworthiness of financial records. Financial audits increasingly rely on machine learning (ML) algorithms to identify irregular or potentially fraudulent journal entries, each characterized by a varying number of transactions. In machine learning, heterogeneity in feature dimensions adds significant complexity to data analysis. In this paper, we introduce a novel approach to anomaly detection in financial data using Large Language Models (LLMs) embeddings. To encode non-semantic categorical data from real-world financial records, we tested 3 pre-trained general purpose sentence-transformer models. For the downstream classification task, we implemented and evaluated 5 optimized ML models including Logistic Regression, Random Forest, Gradient Boosting Machines, Support Vector Machines, and Neural Networks. Our experiments demonstrate that LLMs contribute valuable information to anomaly detection as our models outperform the baselines, in selected settings even by a large margin. The findings further underscore the effectiveness of LLMs in enhancing anomaly detection in financial journal entries, particularly by tackling feature sparsity. We discuss a promising perspective on using LLM embeddings for non-semantic data in the financial context and beyond. |
| Date: | 2024–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2406.03614&r= |
| By: | Edward Sharkey; Philip Treleaven |
| Abstract: | The paper benchmarks several Transformer models [4], to show how these models can judge sentiment from a news event. This signal can then be used for downstream modelling and signal identification for commodity trading. We find that fine-tuned BERT models outperform fine-tuned or vanilla GPT models on this task. Transformer models have revolutionized the field of natural language processing (NLP) in recent years, achieving state-of-the-art results on various tasks such as machine translation, text summarization, question answering, and natural language generation. Among the most prominent transformer models are Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-trained Transformer (GPT), which differ in their architectures and objectives. A CopBERT model training data and process overview is provided. The CopBERT model outperforms similar domain specific BERT trained models such as FinBERT. The below confusion matrices show the performance on CopBERT & CopGPT respectively. We see a ~10 percent increase in f1_score when compare CopBERT vs GPT4 and 16 percent increase vs CopGPT. Whilst GPT4 is dominant It highlights the importance of considering alternatives to GPT models for financial engineering tasks, given risks of hallucinations, and challenges with interpretability. We unsurprisingly see the larger LLMs outperform the BERT models, with predictive power. In summary BERT is partially the new XGboost, what it lacks in predictive power it provides with higher levels of interpretability. Concluding that BERT models might not be the next XGboost [2], but represent an interesting alternative for financial engineering tasks, that require a blend of interpretability and accuracy. |
| Date: | 2024–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2405.12990&r= |
| By: | Tom\'as Aguirre |
| Abstract: | As frontier AI models progress, policy proposals for safe AI development are gaining increasing attention from researchers and policymakers. This paper evaluates the present landscape of integration within the AI supply chain, emphasizing vertical relations and strategic partnerships, with the goal of laying the groundwork to further understand the implications of various governance interventions, including antitrust. The study has two main contributions. First, it maps the AI supply chain by profiling 25 leading companies, examining their 300 pairwise relationships, and noting approximately 80 significant mergers and acquisitions, and 40 relevant antitrust litigation. Second, it offers a conceptual discussion on market definitions and integration drivers, investigating major players like AI labs and chip designers, horizontal integration levels, and vertical relationships among industry leaders. To further understand the strategic partnerships in the industry, we provide three brief case studies, featuring companies such as OpenAI and Nvidia. We conclude by posing open research questions regarding market dynamics and potential governance interventions, such as licensing and safety audits. |
| Date: | 2024–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2406.01722&r= |
| By: | Askitas, Nikos (IZA) |
| Abstract: | This paper addresses the steep learning curve in Machine Learning faced by noncomputer scientists, particularly social scientists, stemming from the absence of a primer on its fundamental principles. I adopt a pedagogical strategy inspired by the adage "once you understand OLS, you can work your way up to any other estimator, " and apply it to Machine Learning. Focusing on a single-hidden-layer artificial neural network, the paper discusses its mathematical underpinnings, including the pivotal Universal Approximation Theorem—an essential "existence theorem". The exposition extends to the algorithmic exploration of solutions, specifically through "feed forward" and "back-propagation", and rounds up with the practical implementation in Python. The objective of this primer is to equip readers with a solid elementary comprehension of first principles and fire some trailblazers to the forefront of AI and causal machine learning. |
| Keywords: | machine learning, deep learning, supervised learning, artificial neural network, perceptron, Python, keras, tensorflow, universal approximation theorem |
| JEL: | C01 C87 C00 C60 |
| Date: | 2024–05 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp17014&r= |