Andreas L. Opdahl, Natali Helberger, Nicholas Diakopoulos
{"title":"特邀社论:人工智能与新闻","authors":"Andreas L. Opdahl, Natali Helberger, Nicholas Diakopoulos","doi":"10.1002/aaai.12179","DOIUrl":null,"url":null,"abstract":"<p>In a time of rising populism and strategic disinformation, quality journalism has become more important than ever. Trusted and high-quality media outlets are needed to provide accurate information to the public in order to protect public safety and wellbeing while supporting the information needs of citizens in order to promote healthy liberal democracies. But quality journalism is also under pressure due to competition for attention from new information channels, declining trust in institutions, and dwindling resources to support the information needs of local communities while there are simultaneously new resource demands to mitigate the impacts of mis- and disinformation. Given this challenging context, how can Artificial Intelligence (AI) support the provision of quality information for society?</p><p>This special issue therefore examines how ongoing advances in AI, including Machine Learning (ML), and generative AI such as Large Language Models (LLMs), can be harnessed to support efficient production and distribution of high-quality news. It takes a broad outlook on the area, including articles that deal with uses and implications of AI in all stages of news production and dissemination, from gathering and analyzing information to creating, presenting, or recommending news content, while also dealing with an onslaught of mis- and disinformation in the broader online information ecosystem. It also discusses AI on different levels, from individual news production tasks, through organizational transformations and ramifications, to societal and economic conditions and consequences. A common red thread throughout the articles is that AI has great transformational potential, also in the media sector, but the factors driving and enabling such transformations are not only technological. Such factors also very much pertain to the broader organizational, infrastructure and economic context, and successful alignment of the different actors along the value chain, including media users.</p><p>The articles presented here offer an optimistic picture of how quality information and the media ecosystem might evolve in positive ways in light of the technological change driven by AI. And while critical approaches and research are by all means warranted such that professional ethical commitments are maintained, we hope this collection at least provides some ideas and inspiration for technologists and other stakeholders to engage further with how to orient their work towards addressing problems, seeking fruitful cooperations with the different stakeholders along the value chain and providing benefits to support quality media production.</p><p>Next, we outline the six articles in the collection providing a brief summary of each to orient the reader.</p><p>LLMs and other generative AI technologies are ushering in a new phase of disruption in the news industry that may affect news production and consumption as well as distribution. David Caswell, in his paper <i>Audiences, Automation and AI: From Structured News to Language Models</i>, argues that a large news organization like the British Broadcasting Corporation (BBC) is prepared for this shift due to previous innovations in automating workflows for personalized content using structured techniques. Such earlier innovations have not only advanced the integration of LLMs but also spurred the development of flexible infrastructures that are resilient in the face of uncertain audience behaviors and editorial processes in AI-driven news environments.</p><p>In the financial news domain, AI is reshaping journalism and fostering a new era of AI-assisted news processes that must be underpinned by trust and accuracy. Claudia Quinonez and Edgar Meij's paper <i>A New Era of AI-Assisted Journalism at Bloomberg</i> provides examples of how Bloomberg has explored AI models for tasks like updated headline generation and controllable text summarization. The authors also discuss automation in Bloomberg's newsroom, where software bots automate story creation for speedier and deeper financial reporting. The paper also examines the broader implications of generative AI in journalism, emphasizing that rigorous standards of accuracy are essential for financial audiences.</p><p>Many local news organizations are exploring uses of AI to address economic pressures and enhance value creation in the face of declining advertising and audience revenues. In their paper <i>Transforming the Value Chain of Local Journalism with Artificial Intelligence</i>, Bartosz Wilczek, Mario Haim, and Neil Thurman present an overview of AI's potential in local news, identifying areas where AI can be most beneficial. They also discuss implementation challenges along the entire value creation chain that are specific to local newsrooms, including resource limitations, and suggest strategies for overcoming them.</p><p>Personalized news recommender systems have become influential in shaping public opinions and decisions. Nava Tintarev, Martijn Willemsen, and Bart P Knijnenburg's paper <i>Measuring the Benefit of Increased Transparency and Control in News Recommendation</i> explains how providing explanations to users can help them understand why certain news items are recommended and enable them to align their reading habits with personal goals, such as knowledge expansion and viewpoint diversity. The authors argue that more realistic evaluations in live recommendation environments are needed to assess the real-world impact of explanatory interventions on user behavior.</p><p>Correcting misinformation involves complex challenges due to psychological, social, and technical factors. In their paper <i>Exploring the Impact of Automated Correction of Misinformation in Social Media</i>, Gregoire Burel, Mohammadali Tavakoli, and Harith Alani argue that the effectiveness of AI-driven corrective methods in real-world settings is under-researched. They examine how misinformation-sharing users have reacted to different types of bot-generated corrective social-media messages, offering new understanding of how corrective messages should be formulated and which types of users to target.</p><p>On the societal level, AI is changing the economic structure and financing of news organizations. In the final paper of this special issue, <i>The Business of News in the AI Economy</i>, Helle Sjøvaag examines the impact of AI on competition, mergers, acquisitions, and IT capabilities in the news industry and discusses how AI influences journalism's traditional business models. The aim is to provide a vocabulary for understanding the economic future of journalism in a data-driven and AI-powered platform economy.</p><p>The authors declare that there is no conflict.</p>","PeriodicalId":7854,"journal":{"name":"Ai Magazine","volume":"45 2","pages":"172-173"},"PeriodicalIF":2.5000,"publicationDate":"2024-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/aaai.12179","citationCount":"0","resultStr":"{\"title\":\"Guest Editorial: AI and the news\",\"authors\":\"Andreas L. Opdahl, Natali Helberger, Nicholas Diakopoulos\",\"doi\":\"10.1002/aaai.12179\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>In a time of rising populism and strategic disinformation, quality journalism has become more important than ever. Trusted and high-quality media outlets are needed to provide accurate information to the public in order to protect public safety and wellbeing while supporting the information needs of citizens in order to promote healthy liberal democracies. But quality journalism is also under pressure due to competition for attention from new information channels, declining trust in institutions, and dwindling resources to support the information needs of local communities while there are simultaneously new resource demands to mitigate the impacts of mis- and disinformation. Given this challenging context, how can Artificial Intelligence (AI) support the provision of quality information for society?</p><p>This special issue therefore examines how ongoing advances in AI, including Machine Learning (ML), and generative AI such as Large Language Models (LLMs), can be harnessed to support efficient production and distribution of high-quality news. It takes a broad outlook on the area, including articles that deal with uses and implications of AI in all stages of news production and dissemination, from gathering and analyzing information to creating, presenting, or recommending news content, while also dealing with an onslaught of mis- and disinformation in the broader online information ecosystem. It also discusses AI on different levels, from individual news production tasks, through organizational transformations and ramifications, to societal and economic conditions and consequences. A common red thread throughout the articles is that AI has great transformational potential, also in the media sector, but the factors driving and enabling such transformations are not only technological. Such factors also very much pertain to the broader organizational, infrastructure and economic context, and successful alignment of the different actors along the value chain, including media users.</p><p>The articles presented here offer an optimistic picture of how quality information and the media ecosystem might evolve in positive ways in light of the technological change driven by AI. And while critical approaches and research are by all means warranted such that professional ethical commitments are maintained, we hope this collection at least provides some ideas and inspiration for technologists and other stakeholders to engage further with how to orient their work towards addressing problems, seeking fruitful cooperations with the different stakeholders along the value chain and providing benefits to support quality media production.</p><p>Next, we outline the six articles in the collection providing a brief summary of each to orient the reader.</p><p>LLMs and other generative AI technologies are ushering in a new phase of disruption in the news industry that may affect news production and consumption as well as distribution. David Caswell, in his paper <i>Audiences, Automation and AI: From Structured News to Language Models</i>, argues that a large news organization like the British Broadcasting Corporation (BBC) is prepared for this shift due to previous innovations in automating workflows for personalized content using structured techniques. Such earlier innovations have not only advanced the integration of LLMs but also spurred the development of flexible infrastructures that are resilient in the face of uncertain audience behaviors and editorial processes in AI-driven news environments.</p><p>In the financial news domain, AI is reshaping journalism and fostering a new era of AI-assisted news processes that must be underpinned by trust and accuracy. Claudia Quinonez and Edgar Meij's paper <i>A New Era of AI-Assisted Journalism at Bloomberg</i> provides examples of how Bloomberg has explored AI models for tasks like updated headline generation and controllable text summarization. The authors also discuss automation in Bloomberg's newsroom, where software bots automate story creation for speedier and deeper financial reporting. The paper also examines the broader implications of generative AI in journalism, emphasizing that rigorous standards of accuracy are essential for financial audiences.</p><p>Many local news organizations are exploring uses of AI to address economic pressures and enhance value creation in the face of declining advertising and audience revenues. In their paper <i>Transforming the Value Chain of Local Journalism with Artificial Intelligence</i>, Bartosz Wilczek, Mario Haim, and Neil Thurman present an overview of AI's potential in local news, identifying areas where AI can be most beneficial. They also discuss implementation challenges along the entire value creation chain that are specific to local newsrooms, including resource limitations, and suggest strategies for overcoming them.</p><p>Personalized news recommender systems have become influential in shaping public opinions and decisions. Nava Tintarev, Martijn Willemsen, and Bart P Knijnenburg's paper <i>Measuring the Benefit of Increased Transparency and Control in News Recommendation</i> explains how providing explanations to users can help them understand why certain news items are recommended and enable them to align their reading habits with personal goals, such as knowledge expansion and viewpoint diversity. The authors argue that more realistic evaluations in live recommendation environments are needed to assess the real-world impact of explanatory interventions on user behavior.</p><p>Correcting misinformation involves complex challenges due to psychological, social, and technical factors. In their paper <i>Exploring the Impact of Automated Correction of Misinformation in Social Media</i>, Gregoire Burel, Mohammadali Tavakoli, and Harith Alani argue that the effectiveness of AI-driven corrective methods in real-world settings is under-researched. They examine how misinformation-sharing users have reacted to different types of bot-generated corrective social-media messages, offering new understanding of how corrective messages should be formulated and which types of users to target.</p><p>On the societal level, AI is changing the economic structure and financing of news organizations. In the final paper of this special issue, <i>The Business of News in the AI Economy</i>, Helle Sjøvaag examines the impact of AI on competition, mergers, acquisitions, and IT capabilities in the news industry and discusses how AI influences journalism's traditional business models. The aim is to provide a vocabulary for understanding the economic future of journalism in a data-driven and AI-powered platform economy.</p><p>The authors declare that there is no conflict.</p>\",\"PeriodicalId\":7854,\"journal\":{\"name\":\"Ai Magazine\",\"volume\":\"45 2\",\"pages\":\"172-173\"},\"PeriodicalIF\":2.5000,\"publicationDate\":\"2024-05-30\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://onlinelibrary.wiley.com/doi/epdf/10.1002/aaai.12179\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Ai Magazine\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://onlinelibrary.wiley.com/doi/10.1002/aaai.12179\",\"RegionNum\":4,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Ai Magazine","FirstCategoryId":"94","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/aaai.12179","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0
摘要
在民粹主义和战略性虚假信息抬头的时代,高质量的新闻报道比以往任何时候都更加重要。我们需要可信赖的高质量媒体为公众提供准确的信息,以保护公众的安全和福祉,同时支持公民的信息需求,以促进健康的自由民主。但是,由于新的信息渠道争夺注意力、对机构的信任度下降、支持当地社区信息需求的资源不断减少,同时为减轻错误信息和虚假信息的影响又提出了新的资源需求,高质量的新闻报道也面临着压力。因此,本特刊将探讨如何利用机器学习(ML)和大型语言模型(LLMs)等生成式人工智能等人工智能领域的最新进展,支持高质量新闻的高效制作和传播。它以广阔的视角审视这一领域,收录的文章涉及人工智能在新闻制作和传播各个阶段的应用和影响,从收集和分析信息到创建、呈现或推荐新闻内容,同时还应对更广泛的在线信息生态系统中大量的错误信息和虚假信息。文章还从不同层面讨论了人工智能,从个人新闻制作任务,到组织变革和影响,再到社会和经济条件及后果。整篇文章的一条共同主线是,人工智能具有巨大的变革潜力,在媒体领域也是如此,但推动和促成这种变革的因素不仅仅是技术。这里介绍的文章乐观地描绘了在人工智能技术变革的推动下,优质信息和媒体生态系统如何以积极的方式发展。我们希望本文集至少能为技术专家和其他利益相关者提供一些想法和启发,使他们能进一步思考如何将自己的工作导向解决问题、寻求与价值链上不同利益相关者的富有成效的合作,并为支持优质媒体生产提供益处。接下来,我们将概述文集中的六篇文章,并对每篇文章进行简要概述,以便为读者指明方向。LLM 和其他生成式人工智能技术正在为新闻行业带来一个新的颠覆阶段,可能会影响新闻生产、消费和传播。大卫-卡斯维尔(David Caswell)在他的论文《受众、自动化与人工智能:从结构化新闻到语言模型》中认为,像英国广播公司(BBC)这样的大型新闻机构已经为这种转变做好了准备,因为之前已经利用结构化技术实现了个性化内容自动化工作流程的创新。在财经新闻领域,人工智能正在重塑新闻业,并促进人工智能辅助新闻流程进入一个新时代,而这必须以信任和准确性为基础。Claudia Quinonez 和 Edgar Meij 的论文《A New Era of AI-Assisted Journalism at Bloomberg》(彭博社的人工智能辅助新闻新时代)举例说明了彭博社如何在更新标题生成和可控文本摘要等任务中探索人工智能模型。作者还讨论了彭博社新闻编辑室的自动化问题,在那里,软件机器人自动进行新闻创作,以实现更快速、更深入的财经报道。论文还探讨了生成式人工智能对新闻业的广泛影响,强调严格的准确性标准对财经受众至关重要。面对广告和受众收入的下降,许多地方新闻机构正在探索如何利用人工智能来应对经济压力,提高价值创造能力。Bartosz Wilczek、Mario Haim 和 Neil Thurman 在他们的论文《用人工智能改造地方新闻价值链》中概述了人工智能在地方新闻中的潜力,指出了人工智能最能带来益处的领域。他们还讨论了地方新闻编辑室在整个价值创造链中面临的具体实施挑战,包括资源限制,并提出了克服这些挑战的策略。 Nava Tintarev、Martijn Willemsen 和 Bart P Knijnenburg 的论文《衡量在新闻推荐中增加透明度和控制的益处》解释了向用户提供解释如何帮助他们理解某些新闻条目被推荐的原因,并使他们的阅读习惯符合个人目标,如知识扩展和观点多样性。作者认为,需要在实时推荐环境中进行更真实的评估,以评估解释性干预对用户行为的实际影响。由于心理、社会和技术因素,纠正错误信息涉及复杂的挑战。Gregoire Burel、Mohammadali Tavakoli 和 Harith Alani 在论文《探索自动纠正社交媒体中错误信息的影响》中指出,人工智能驱动的纠正方法在现实环境中的有效性还没有得到充分研究。他们研究了分享错误信息的用户对不同类型的机器人生成的社交媒体纠正信息的反应,为如何制定纠正信息以及针对哪类用户提供了新的理解。在社会层面,人工智能正在改变经济结构和新闻机构的融资方式。在本特刊的最后一篇论文《人工智能经济中的新闻业务》中,Helle Sjøvaag 探讨了人工智能对新闻行业竞争、并购和 IT 能力的影响,并讨论了人工智能如何影响新闻业的传统业务模式。其目的是为理解新闻业在数据驱动和人工智能推动的平台经济中的经济前景提供一个词汇表。
In a time of rising populism and strategic disinformation, quality journalism has become more important than ever. Trusted and high-quality media outlets are needed to provide accurate information to the public in order to protect public safety and wellbeing while supporting the information needs of citizens in order to promote healthy liberal democracies. But quality journalism is also under pressure due to competition for attention from new information channels, declining trust in institutions, and dwindling resources to support the information needs of local communities while there are simultaneously new resource demands to mitigate the impacts of mis- and disinformation. Given this challenging context, how can Artificial Intelligence (AI) support the provision of quality information for society?
This special issue therefore examines how ongoing advances in AI, including Machine Learning (ML), and generative AI such as Large Language Models (LLMs), can be harnessed to support efficient production and distribution of high-quality news. It takes a broad outlook on the area, including articles that deal with uses and implications of AI in all stages of news production and dissemination, from gathering and analyzing information to creating, presenting, or recommending news content, while also dealing with an onslaught of mis- and disinformation in the broader online information ecosystem. It also discusses AI on different levels, from individual news production tasks, through organizational transformations and ramifications, to societal and economic conditions and consequences. A common red thread throughout the articles is that AI has great transformational potential, also in the media sector, but the factors driving and enabling such transformations are not only technological. Such factors also very much pertain to the broader organizational, infrastructure and economic context, and successful alignment of the different actors along the value chain, including media users.
The articles presented here offer an optimistic picture of how quality information and the media ecosystem might evolve in positive ways in light of the technological change driven by AI. And while critical approaches and research are by all means warranted such that professional ethical commitments are maintained, we hope this collection at least provides some ideas and inspiration for technologists and other stakeholders to engage further with how to orient their work towards addressing problems, seeking fruitful cooperations with the different stakeholders along the value chain and providing benefits to support quality media production.
Next, we outline the six articles in the collection providing a brief summary of each to orient the reader.
LLMs and other generative AI technologies are ushering in a new phase of disruption in the news industry that may affect news production and consumption as well as distribution. David Caswell, in his paper Audiences, Automation and AI: From Structured News to Language Models, argues that a large news organization like the British Broadcasting Corporation (BBC) is prepared for this shift due to previous innovations in automating workflows for personalized content using structured techniques. Such earlier innovations have not only advanced the integration of LLMs but also spurred the development of flexible infrastructures that are resilient in the face of uncertain audience behaviors and editorial processes in AI-driven news environments.
In the financial news domain, AI is reshaping journalism and fostering a new era of AI-assisted news processes that must be underpinned by trust and accuracy. Claudia Quinonez and Edgar Meij's paper A New Era of AI-Assisted Journalism at Bloomberg provides examples of how Bloomberg has explored AI models for tasks like updated headline generation and controllable text summarization. The authors also discuss automation in Bloomberg's newsroom, where software bots automate story creation for speedier and deeper financial reporting. The paper also examines the broader implications of generative AI in journalism, emphasizing that rigorous standards of accuracy are essential for financial audiences.
Many local news organizations are exploring uses of AI to address economic pressures and enhance value creation in the face of declining advertising and audience revenues. In their paper Transforming the Value Chain of Local Journalism with Artificial Intelligence, Bartosz Wilczek, Mario Haim, and Neil Thurman present an overview of AI's potential in local news, identifying areas where AI can be most beneficial. They also discuss implementation challenges along the entire value creation chain that are specific to local newsrooms, including resource limitations, and suggest strategies for overcoming them.
Personalized news recommender systems have become influential in shaping public opinions and decisions. Nava Tintarev, Martijn Willemsen, and Bart P Knijnenburg's paper Measuring the Benefit of Increased Transparency and Control in News Recommendation explains how providing explanations to users can help them understand why certain news items are recommended and enable them to align their reading habits with personal goals, such as knowledge expansion and viewpoint diversity. The authors argue that more realistic evaluations in live recommendation environments are needed to assess the real-world impact of explanatory interventions on user behavior.
Correcting misinformation involves complex challenges due to psychological, social, and technical factors. In their paper Exploring the Impact of Automated Correction of Misinformation in Social Media, Gregoire Burel, Mohammadali Tavakoli, and Harith Alani argue that the effectiveness of AI-driven corrective methods in real-world settings is under-researched. They examine how misinformation-sharing users have reacted to different types of bot-generated corrective social-media messages, offering new understanding of how corrective messages should be formulated and which types of users to target.
On the societal level, AI is changing the economic structure and financing of news organizations. In the final paper of this special issue, The Business of News in the AI Economy, Helle Sjøvaag examines the impact of AI on competition, mergers, acquisitions, and IT capabilities in the news industry and discusses how AI influences journalism's traditional business models. The aim is to provide a vocabulary for understanding the economic future of journalism in a data-driven and AI-powered platform economy.
期刊介绍:
AI Magazine publishes original articles that are reasonably self-contained and aimed at a broad spectrum of the AI community. Technical content should be kept to a minimum. In general, the magazine does not publish articles that have been published elsewhere in whole or in part. The magazine welcomes the contribution of articles on the theory and practice of AI as well as general survey articles, tutorial articles on timely topics, conference or symposia or workshop reports, and timely columns on topics of interest to AI scientists.