|
on Artificial Intelligence |
| By: | Paul Duetting; Vahab Mirrokni; Renato Paes Leme; Haifeng Xu; Song Zuo |
| Abstract: | We investigate auction mechanisms to support the emerging format of AI-generated content. We in particular study how to aggregate several LLMs in an incentive compatible manner. In this problem, the preferences of each agent over stochastically generated contents are described/encoded as an LLM. A key motivation is to design an auction format for AI-generated ad creatives to combine inputs from different advertisers. We argue that this problem, while generally falling under the umbrella of mechanism design, has several unique features. We propose a general formalism -- the token auction model -- for studying this problem. A key feature of this model is that it acts on a token-by-token basis and lets LLM agents influence generated contents through single dimensional bids. We first explore a robust auction design approach, in which all we assume is that agent preferences entail partial orders over outcome distributions. We formulate two natural incentive properties, and show that these are equivalent to a monotonicity condition on distribution aggregation. We also show that for such aggregation functions, it is possible to design a second-price auction, despite the absence of bidder valuation functions. We then move to designing concrete aggregation functions by focusing on specific valuation forms based on KL-divergence, a commonly used loss function in LLM. The welfare-maximizing aggregation rules turn out to be the weighted (log-space) convex combination of the target distributions from all participants. We conclude with experimental results in support of the token auction formulation. |
| Date: | 2023–10 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2310.10826&r=ain |
| By: | Oldenburg, Derk; Serentschy, Georg |
| Abstract: | Regulating emerging technologies has always been a challenge. Balancing the risks and opportunities of AI is raising the bar, in particular since generative AI systems such as ChatGPT have been made accessible to a large public. The rapid development of AI deserves another look at what balancing means and how we could look at it in the face of the challenges confronting our society. |
| Date: | 2023 |
| URL: | http://d.repec.org/n?u=RePEc:zbw:itse23:278007&r=ain |
| By: | Susan Ariel Aaronson (George Washington University) |
| Abstract: | Only 8 months have passed since Chat-GPT and the large learning model underpinning it took the world by storm. This article focuses on the data supply chain-the data collected and then utilized to train large language models and the governance challenge it presents to policymakers. These challenges include: - How web scraping may affect individuals and firms which hold copyrights. - How web scraping may affect individuals and groups who are supposed to be protected under privacy and personal data protection laws. - How web scraping revealed the lack of protections for content creators and content providers on open access web sites; and - How the debate over open and closed source LLM reveals the lack of clear and universal rules to ensure the quality and validity of datasets. As the US National Institute of Standards explained, many LLMs depend on "largescale datasets, which can lead to data quality and validity concerns. "The difficulty of finding the "right" data may lead AI actors to select datasets based more on accessibility and availability than on suitability... Such decisions could contribute to an environment where the data used in processes is not fully representative of the populations or phenomena that are being modeled, introducing downstream risks" -in short problems of quality and validity (NIST: 2023, 80). Thie author uses qualitative methods to examine these data governance challenges. In general, this report discusses only those governments that adopted specific steps (actions, policies, new regulations etc.) to address web scraping, LLMs, or generative AI. The author acknowledges that these examples do not comprise a representative sample based on income, LLM expertise, and geographic diversity. However, the author uses these examples to show that while some policymakers are responsive to rising concerns, they do not seem to be looking at these issues systemically. A systemic approach has two components: First policymakers recognize that these AI chatbots are a complex system with different sources of data, that are linked to other systems designed, developed, owned, and controlled by different people and organizations. Data and algorithm production, deployment, and use are distributed among a wide range of actors who together produce the system's outcomes and functionality. Hence accountability is diffused and opaque(Cobbe et al: 2023). Secondly, as a report for the US National Academy of Sciences notes, the only way to govern such complex systems is to create "a governance ecosystem that cuts across sectors and disciplinary silos and solicits and addresses the concerns of many stakeholders." This assessment is particularly true for LLMs—a global product with a global supply chain with numerous interdependencies among those who supply data, those who control data, and those who are data subjects or content creators (Cobbe et al: 2023). |
| Keywords: | data, data governance, personal data, property rights, open data, open source, governance |
| JEL: | P51 |
| Date: | 2023–07 |
| URL: | http://d.repec.org/n?u=RePEc:gwi:wpaper:2023-07&r=ain |
| By: | Mariam Dundua (Financial and Supervisory Technology Development Department, National Bank of Georgia); Otar Gorgodze (Head of Financial and Supervisory Technologies Department, National Bank of Georgia) |
| Abstract: | The recent advances in Artificial Intelligence (AI), in particular, the development of reinforcement learning (RL) methods, are specifically suited for application to complex economic problems. We formulate a new approach looking for optimal monetary policy rules using RL. Analysis of AI generated monetary policy rules indicates that optimal policy rules exhibit significant nonlinearities. This could explain why simple monetary rules based on traditional linear modeling toolkits lack the robustness needed for practical application. The generated transition equations analysis allows us to estimate the neutral policy rate, which came out to be 6.5 percent. We discuss the potential combination of the method with state-of-the-art FinTech developments in digital finance like DeFi and CBDC and the feasibility of MonetaryTech approach to monetary policy. |
| Keywords: | Artificial Intelligence; Reinforcement Learning; Monetary policy |
| JEL: | C60 C61 C63 E17 C45 E52 |
| Date: | 2022–11 |
| URL: | http://d.repec.org/n?u=RePEc:aez:wpaper:02/2022&r=ain |
| By: | Sam Hainsworth |
| Abstract: | Media hype and technological breakthroughs are fuelling the race to adopt Artificial Intelligence amongst the business community, but is there evidence to suggest this will increase productivity? This paper uses 2015-2019 microdata from the UK Office for National Statistics to identify if the adoption of Artificial Intelligence techniques increases labour productivity in UK businesses. Using fixed effects estimation (Within Group) with a log-linear regression specification the paper concludes that there is no statistically significant impact of AI adoption on labour productivity. |
| Date: | 2023–10 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2310.05985&r=ain |
| By: | Kauhanen, Antti; Pajarinen, Mika; Rouvinen, Petri |
| Abstract: | Abstract About 19% of Finnish employment is in occupations with at least 50% of tasks exposed to Generative Artificial Intelligence (GAI) with text- and code-generating abilities, such as ChatGPT. Most jobs need some adjustment due to recent advances in GAI, but relatively few will be heavily disrupted. Our results do not support the “end-of-work” narrative. GAI’s long-term impact on human employment is ambiguous; its effects could certainly be positive, especially if GAI turns out to be a sustained source of productivity growth. Whatever the outcome, our findings suggest that a labor market change induced by GAI is brewing and that individuals, organizations, and society all need to make a conscious decision to adapt. In our view, the biggest risk of GAI in the Finnish labor market is that we will not explore the opportunities it offers with any enthusiasm. Its impact is best faced head-on, and early adopters stand to benefit the most from it. More broadly, the biggest societal risk – in our view – is that we are less and less capable of separating human- and GAI-generated digital content (including audio, images, and video), with a heightened risk of disinformation and highly targeted cyber-attacks. This research brief replicates the analysis by Eloundou et al. (2023) in the context of Finland. |
| Keywords: | Generative artificial intelligence, Technological change, Employment, Labor market, Occupations |
| JEL: | E24 J21 O33 |
| Date: | 2023–10–25 |
| URL: | http://d.repec.org/n?u=RePEc:rif:briefs:127&r=ain |
| By: | Kauhanen, Antti; Pajarinen, Mika; Rouvinen, Petri |
| Abstract: | Abstract Some one-fifth of Finnish employment is in occupations with at least half of tasks exposed to generative artificial intelligence. A relatively large share of occupations has at least some exposure, but few occupations have high exposures. Contrary to prior technological discontinuities, in the case of generative artificial intelligence the labor market elite is relatively more exposed. As far as the Finnish labor market is concerned, the effect of generative artificial intelligence is ambiguous – and quite possibly positive. Regardless, employees are faced with a sizable change, which is best addressed head-on, i.e., by experimenting with and deploying generative artificial intelligence as soon as possible. Our observations are based on a replication of the US analysis by Eloundou et al. (2023) in the context of Finland. This brief kicks off a research project conducted by ETLA and supported by the TT foundation. |
| Keywords: | Generative artificial intelligence, Technological change, Employment, Labor market, Occupations |
| JEL: | E24 J21 O33 |
| Date: | 2023–10–25 |
| URL: | http://d.repec.org/n?u=RePEc:rif:briefs:128&r=ain |
| By: | Yujie Ding; Shuai Jia; Tianyi Ma; Bingcheng Mao; Xiuze Zhou; Liuliu Li; Dongming Han |
| Abstract: | The remarkable achievements and rapid advancements of Large Language Models (LLMs) such as ChatGPT and GPT-4 have showcased their immense potential in quantitative investment. Traders can effectively leverage these LLMs to analyze financial news and predict stock returns accurately. However, integrating LLMs into existing quantitative models presents two primary challenges: the insufficient utilization of semantic information embedded within LLMs and the difficulties in aligning the latent information within LLMs with pre-existing quantitative stock features. We propose a novel framework consisting of two components to surmount these challenges. The first component, the Local-Global (LG) model, introduces three distinct strategies for modeling global information. These approaches are grounded respectively on stock features, the capabilities of LLMs, and a hybrid method combining the two paradigms. The second component, Self-Correlated Reinforcement Learning (SCRL), focuses on aligning the embeddings of financial news generated by LLMs with stock features within the same semantic space. By implementing our framework, we have demonstrated superior performance in Rank Information Coefficient and returns, particularly compared to models relying only on stock features in the China A-share market. |
| Date: | 2023–10 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2310.05627&r=ain |
| By: | Neng Wang; Hongyang Yang; Christina Dan Wang |
| Abstract: | In the swiftly expanding domain of Natural Language Processing (NLP), the potential of GPT-based models for the financial sector is increasingly evident. However, the integration of these models with financial datasets presents challenges, notably in determining their adeptness and relevance. This paper introduces a distinctive approach anchored in the Instruction Tuning paradigm for open-source large language models, specifically adapted for financial contexts. Through this methodology, we capitalize on the interoperability of open-source models, ensuring a seamless and transparent integration. We begin by explaining the Instruction Tuning paradigm, highlighting its effectiveness for immediate integration. The paper presents a benchmarking scheme designed for end-to-end training and testing, employing a cost-effective progression. Firstly, we assess basic competencies and fundamental tasks, such as Named Entity Recognition (NER) and sentiment analysis to enhance specialization. Next, we delve into a comprehensive model, executing multi-task operations by amalgamating all instructional tunings to examine versatility. Finally, we explore the zero-shot capabilities by earmarking unseen tasks and incorporating novel datasets to understand adaptability in uncharted terrains. Such a paradigm fortifies the principles of openness and reproducibility, laying a robust foundation for future investigations in open-source financial large language models (FinLLMs). |
| Date: | 2023–10 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2310.04793&r=ain |
| By: | Boyu Zhang; Hongyang Yang; Tianyu Zhou; Ali Babar; Xiao-Yang Liu |
| Abstract: | Financial sentiment analysis is critical for valuation and investment decision-making. Traditional NLP models, however, are limited by their parameter size and the scope of their training datasets, which hampers their generalization capabilities and effectiveness in this field. Recently, Large Language Models (LLMs) pre-trained on extensive corpora have demonstrated superior performance across various NLP tasks due to their commendable zero-shot abilities. Yet, directly applying LLMs to financial sentiment analysis presents challenges: The discrepancy between the pre-training objective of LLMs and predicting the sentiment label can compromise their predictive performance. Furthermore, the succinct nature of financial news, often devoid of sufficient context, can significantly diminish the reliability of LLMs' sentiment analysis. To address these challenges, we introduce a retrieval-augmented LLMs framework for financial sentiment analysis. This framework includes an instruction-tuned LLMs module, which ensures LLMs behave as predictors of sentiment labels, and a retrieval-augmentation module which retrieves additional context from reliable external sources. Benchmarked against traditional models and LLMs like ChatGPT and LLaMA, our approach achieves 15\% to 48\% performance gain in accuracy and F1 score. |
| Date: | 2023–10 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2310.04027&r=ain |
| By: | Saeid Vaghefi (University of Zurich); Qian Wang (University of Zurich; Inovest Partners AG); Veruska Muccione (University of Zurich; University of Geneva); Jingwei Ni (ETH Zurich); Mathias Kraus (University of Erlangen); Julia Bingler (University of Oxford); Tobias Schimanski (University of Zurich); Chiara Colesanti Senni (ETH Zurich; University of Zurich); Nicolas Webersinke (Friedrich-Alexander-Universität Erlangen-Nürnberg); Christian Huggel (University of Zurich); Markus Leippold (University of Zurich; Swiss Finance Institute) |
| Abstract: | Large Language Models (LLMs) have made significant progress in recent years, achieving remarkable results in question-answering tasks (QA). However, they still face two major challenges: hallucination and outdated information after the training phase. These challenges take center stage in critical domains like climate change, where obtaining accurate and up-to-date information from reliable sources in a limited time is essential and difficult. To overcome these barriers, one potential solution is to provide LLMs with access to external, scientifically accurate, and robust sources (long-term memory) to continuously update their knowledge and prevent the propagation of inaccurate, incorrect, or outdated information. In this study, we enhanced GPT-4 by integrating the information from the Sixth Assessment Report of the Intergovernmental (IPCC AR6), the most comprehensive, up-to-date, and reliable source in this domain. We present our conversational AI prototype, available at www.chatclimate.ai, for his invaluable and voluntary support in setting up the server. The server will become available by mid-April.} and demonstrate its ability to answer challenging questions accurately. The answers and their sources were evaluated by our team of IPCC authors, who used their expert knowledge to score the accuracy of the answers from 1 (very-low) to 5 (very-high). The evaluation showed that the hybrid chatClimate provided more accurate answers, highlighting the effectiveness of our solution. This approach can be easily scaled for chatbots in specific domains, enabling the delivery of reliable and accurate information. |
| Date: | 2023–10 |
| URL: | http://d.repec.org/n?u=RePEc:chf:rpseri:rp2388&r=ain |
| By: | Abdul Latif Baydoun (LUMEN - Lille University Management Lab - ULR 4999 - Université de Lille) |
| Abstract: | AI drives history-based evolution of societies, and we thrive to understand AI's power to empower humanity and nature. Using an ethnographic approach, we try to learn how AI shapes consumer ethical agency considering the important issue of climate change at stake. |
| Date: | 2022–06–13 |
| URL: | http://d.repec.org/n?u=RePEc:hal:journl:hal-04214774&r=ain |