风险防范下算法的监管路径研究

人工智能时代,由源源不断的数据驱动的算法已经成为新的权力代理人。然而,对算法的监管仍依照传统路径,在结果监管的思路指导下,进行事后的内容审查,由网络平台承担法律责任。机器深度学习和自动决策技术的发展超越了结果监管思路的制度假设,给传统监管路径带来严重冲击。应及时树立以风险防范为目的的监管思路,实行内容与算法并重的双轨审查机制,以及设立平台责任与技术责任双轨并行的责任体系,并对算法的生产性资源数据的收集和使用进行合理限制。     

人工智能在金融领域的应用场景和现状分析

随着人工智能的不断发展,人们也在不断思考如何将人工智能技术应用于金融领域,给金融的发展带来实质性的促进,目前已有很多新兴的互联网金融公司诞生,这些公司致力于将金融科技融入到实际产品中,给公司和客户带来收益。那么对于传统的金融公司而言,如何在这场技术变革中,利用好最新技术,将这些技术落实到实际的应用场景中,给业务带来提升和帮助,将会是一个重要的研究课题。本文从图像处理、语音识别、自然语言处理等多个人工智能技术点进行探讨,详细分析这些技术在金融领域可能的应用场景以及目前这些场景的发展状况。     

广发智慧 御风而行——论人工智能在金融领域的应用

人工智能是一门综合了大数据、机器学习以及各类感知智能等的跨学科、跨领域技术,未来将会给各行备业带来巨大的变革。预计随着人工智能的深入发展和应用,金融业将在三个方面受到影响：一是企业运营的大量人工将逐步被机器取代;二是金融服务的广度和深度将极大扩展;三是金融行业获利模式将产生巨大变化。     

人工智能浪潮下财会行业的发展趋势

随着人工智能技术的快速发展,其已被应用到多个行业和领域中,在这股浪潮中,财会行业中会计、税务、审计等多个工作领域也不例外的初步应用了人工智能技术,人工智能明显提升了财务信息的质量、促进财务信息呈现不同方式的变化、有效提高了财务工作的效率和促进了财会行业人才需求类型发生转变,因此,其对财务行业发展具有重要的促进作用,但是鉴于我国目前的人工智能技术发展尚不成熟以及相关财务人员对人工智能认识并不充足,为此,还需不断完善人工智能在财会行业中的发展.本文主要分析了人工智能在财会行业中的应用现状、人工智能给财会行业带来的重要影响以及探究了人工智能浪潮下财会行业的发展趋势.     

中小银行如何用好AI大模型技术——以AI营销助手为例

<正>AI大模型技术相较于过往的ABCD(人工智能、区块链、云计算、大数据)等新技术更加进步，是一场颠覆式的技术变革。AI大模型技术的发展提高了人工智能的普适性和泛化性，使得人工智能可以在更多的领域和场景中应用，从而扩大了数字创新的范围和深度。大模型让我们能够完成一些以前难以想象的事情，如让AI生成图像、诗歌、文本、音乐、编写代码等，也让我们面临一些前所未有的挑战，如数据安全、模型可信、知识产权、人工智能伦理等方面的问题。AI大模型技术不仅改变了处理数据和信息的方式，也改变了我们认识世界和判断事物的方式。AI大模型技术不仅对劳动密集型的工作产生影响，也对脑力密集型的工作产生了深刻影响。未来几年，AI大模型技术逐步成熟，将会重构现有的社会分工以及商业模式，给整个社会带来巨大变化。我们需要在充分利用其优势的同时，注意防范其潜在风险，并积极探索适应其变革的策略和路径，把握好技术变革带来的机遇。     

基于会计目标角度认识人工智能对会计工作的影响

随着科技的迅速发展,人工智能技术已经逐渐渗透到各个行业中,给传统的工作方式带来了革命性的变革。会计,作为企业的核心支撑部门,其工作方式和目标也受到了人工智能的深刻影响。传统的会计工作模式在人工智能的推动下,正面临着极大的挑战与机遇。如何利用人工智能技术提高会计工作的效率和质量,已经成为业界关注的焦点。因此,本文就基于会计目标的角度,深入探讨了人工智能对会计工作的影响,以期为会计行业在人工智能时代的转型与发展提供有益的参考和建议。     

基于TPACK模型的《统计学》课程混合式教学改革研究

此前,教育部正式提出了《教育信息化2.0行动计划》(2018),认为人工智能、区块链、大数据技术的快速发展将会给教育形态带来意义深远的影响和变化,同时提出我国高校教育信息技术与学科教学深度融合不够.现阶段很多高校没有正确认识和应用信息技术,本文基于统计学课程,针对传统统计学课程中存在的问题,应用TPACK理论基础进行混...     

人工智能赋能绿色金融高质量发展的途径和面临的挑战

近些年,由于人工智能技术的不断发展,金融行业可以通过人工智能提高大数据的分析能力,辅助绿色金融的投资,评估企业的绿色金融各项指标,抓取企业环境、社会和治理（ESG）报告的关键数据。金融是强监管的行业,人工智能技术的一些特性使得其在绿色金融领域的应用存在一定风险,其应用过程中给金融监管来带来挑战。本文分析了人工智能技术在促进绿色金融行业高质量发展的作用途径以及面临的挑战,提出应强化人工智能技术重要环节的监控、提高机器学习的可解释性、充分发挥人工智能技术的优势等建议。     

人工智能在税务领域应用中的风险及规制北大核心

人工智能等现代信息技术在提高税收征管效率的同时,也可能面临技术赋能、技术应用和技术伦理方面的困境。因此,有必要对人工智能在税务领域的应用予以法律规制。考察人工智能在税务领域应用中风险的形成机制,我国亟须划定人工智能技术在税务领域的适用边界并完善涉税数据和算法的管理机制,具体包括数据治理、算法透明以及应用边界等方面的制度构建,以实现人工智能和依法治税的深度融合。     

趋利避害 防范新技术风险

近年来,金融科技日新月异,特别是大数据、云计算、人工智能、生物识别、区块链等新技术的不断涌现,实现了信息科技与金融业务的深度融合,推动了金融业务流程与商业模式的变革,在驱动业务数字化、打造智能商业银行、丰富认证支付场景、助力数据治理工作等方面发挥了重要作用。但同时,我们还要清醒地认识到,事物存在两面性,任何一项新的技术都有其局限和不断发展完善的过程,如果不能正确认识并加以防范,就会给我们带来不可预知的风险和损失。     

Almanac of anticipations: The age of affluence—of anticipation—of anxiety—of automation—of discontinuity—of longing—of overkill—of space exploration

With this survey of developments in the futures field since 1945, I.F. Clarke completes his present series for Futures. He makes two major points: that forecasting techniques have become essential tools in the management of change; and that nuclear weapons and ecological problems have added a moral dimension to the study of the future. In the next issue of Futures he will start a new series on the range of future-thinking in the USA, from the expectations of the first settlers to the Star Wars programme.     

社会主义所有制形式的改革、完善与共产党执政能力的提高

在中国,所有制形式的调整、改革与完善,先进生产力的发展,党的执政能力建设,是社会主义建设过程中密切结合在一起的三个不同层面的发展关系.应在相互促进和共同提高中促进社会的和谐发展,保证社会主义现代化建设战略目标的顺利实现.     

基金公司经营融资融券业务的问题审视

融资融券业务正式运营已经开展,为证券市场带来了革命性的改革,作为证券市场的一大主体,基金公司面对融资融券也迎来了新的发展方向和挑战,本文从融资融券业务的运行机制入手,分析了基金公司融资融券的新契机,并对其即将面临的问题进行审视和剖析,进而提出策略建议.     

对基层人民银行科技工作的几点建议

银监会分设后,人民银行将专司货币政策、金融稳定、金融服务三大职能。作为人民银行的分支机构,基层人民银行的工作重心也将从金融监管调整到这三大职能上来。人民银行的科技部门,就要积极运用科技手段,为人民银行履行新的职能发挥“服务、支持、促进、发展”的作用。一、围绕三大职能搭建五大平台经过“六五打基础、七五作准备、八五大发展”几个阶段后,人民银行相继实现了核算电算化、清算网络化和办公信息化。特别是近一两年来,陆续开通了大额支付系统、信贷登记咨询系统、会计四集中系统、金融信息服务网站等系统,人民银行信息化水平又…     

Properties of time-series estimates of degree of leverage measures

Empirical studies suggest that time-series regression estimates of the degrees of operating and financial leverage have a tendency to produce measures less than one. According to ex ante theory, these measures should be greater than one for firms operating above the breakeven point. There have also been suggestions that the biases in these estimates may be attributable to an underlying increase in unit sales. This work presents evidence that these counter-intuitive measures are produced by changes in the firm's operating parameters (unit price, variable cost, fixed cost and interest payments). It further suggests that attempts to control for the underlying change in unit sales substantially increase the volatility of predicted estimates.     

浅议金融机构的举证责任

金融纠纷案件是当事人以存单或进账单、对账单、存款合同、保险单、票据、证券等凭证为主要证据向人民法院提出诉讼的案件,对于不同种类金融纠纷的当事人的举证责任,我国现行的金融法律、法规及司法解释有不同的规定。由于金融机构的特殊法律地位,发生金融纠纷时存在有关当事人将最后偿付风险转移给金融机构承担的主观故意,在客观上会使金融机构难以提供相应的证据来对抗不法债权人的诉讼请求,而导致金融机构败诉。为此,对金融诉讼案件中金融机构的举证责任加以研究具有重要的实践意义。     

Summary of Results of Survey of Seminar Attendees

Abstract

The Society of Actuaries undertook a three-phase research project on mortality improvement in the three NAFTA countries: Canada, Mexico, and the U.S. Phase 1 consisted of a literature review of papers on projecting mortality levels in the future and a study of the trend in mortality improvement during this century. Phase 2 consisted of a discussion of different facets of modeling mortality rates at a seminar attended by 79 experts (actuaries, demographers, economists, and medical researchers) representing different countries. The last session of the seminar consisted of the completion of a survey by the attendees to obtain input for Phase 3, which would analyze the impact of mortality improvement on the social security system of each country. This paper summarizes the results of the survey.

The survey results illustrate the difficulty in forecasting mortality levels, because the effects of many factors that could have significant impact on mortality rates are unknown. This suggests the need for dynamic forecasting, which allows for the possibility of random shocks. A majority of the survey respondents believe that stochastic forecasting models, despite their complexity, have significant potential to add value. Respondents also believe that both historical data and cause-specific mortality forecasts are useful as input and also in validating forecasts of the aggregate levels of mortality. The challenge is to develop more sophisticated forecasting models to produce results that are relatively easy to interpret and to communicate these results to the desired audiences, including the public and policymakers.

The survey results suggest that the aggregate effect of lifestyle changes, medical advances, diseases, catastrophe, and physical environmental changes is an increase in life span. However, there is much uncertainty about the future. Respondents expect that beyond the year 2020 the mean annual rate of reduction in mortality for males age 65 and over will average about 0.58% for Canada, 0.76% for Mexico, and 0.67% for the U.S. The results for the female age 65 and over population are 0.64%, 0.83%, and 0.70%, respectively. The age 65 and over population is expected to see larger percentage reductions in mortality than the 0–14 and 15–64 populations. The reductions in male and female mortality will be ultimately the same, and the mortality levels in the three countries will ultimately converge, although differences may persist for decades.     

The Usefulness of Measures of Consistency of Discretionary Components of Accruals in the Detection of Earnings Management

Abstract:  Prior research has shown the prevalence of measurement error in models used to estimate aggregate discretionary accruals. In these models, the incremental information content of the various components of accruals is ignored. Limited prior research and data gathered from firms under Securities and Exchange Commission (SEC) litigation indicate that managers use either one or more than one component of accruals simultaneously, in a consistent way to manipulate bottom-line earnings in a given direction. I propose two measures that capture the consistency between the discretionary components of accruals and test their significance in earnings management (EM) detection in firms that have artificially added accrual manipulation and firms that were targeted by the SEC for accrual manipulation. There is evidence that this information is incrementally useful in detecting EM. This finding paves the way for improvements in the discretionary accruals measure by including consistency information from the components of aggregate accruals.     

Separation of Unit of Account from Medium of Exchange

The separation of a unit of account (UoA) from a medium of exchange (MoE) in the commodity–money system is investigated by considering explicitly a seller's choice of UoA in terms of either an MoE or a unit of metal weight. If the likelihood of debasement of an MoE and its rate are high enough, the price is posted in terms of a unit of metal weight rather than an MoE. Interestingly, this MoE–UoA separated equilibrium yields the flexible nominal price, whereas an MoE–UoA integrated equilibrium yields the sticky one. This implies the nominal price rigidity in the fiat‐money system where MoE and UoA are integrated.