Generative AI in Journalism

Generative AI Overview

Generative AI refers to artificial intelligence systems that can create new content, such as text, images, and music. It is based on machine learning, particularly deep learning techniques involving neural networks. These neural networks are trained on large datasets to generate outputs that mimic the training data. For example, in text generation, models like GPT (Generative Pretrained Transformer) analyze vast amounts of text to produce coherent and contextually relevant content. Similarly, in image generation, models like Generative Adversarial Networks (GANs) create realistic images by learning from a dataset of existing images. In other words, they are a “class of very powerful AI models that can be used as the basis for other models: they can be specialized, or retrained, or otherwise modified for specific applications” (Loukides, 2023, p. 2).

The transformative potential of generative AI as a technological force is profound, promising to redefine creativity, enhance productivity, and reshape ethical norms. This technology, built on advanced machine learning algorithms and neural networks, is not only automating tasks but also creating novel content, thus challenging our traditional understanding of human creativity.

 

Historical Context

Early AI Development: The evolution of AI from its inception in the mid-20th century includes several key milestones. Alan Turing's computational theories, especially his concept of the Turing Test (Turing, 1950, pp. 433-460) laid foundational ideas for AI. The development of neural networks, a significant leap in AI, began with the perceptron in the 1950s. Over the decades, advancements in computational power and data availability have led to more sophisticated neural networks, enabling modern AI capabilities.

Early technological advancements like the internet and personal computing have significantly influenced societal structures, cultural norms, and economic models (Castells, p. 1). The advent of the internet enabled unprecedented connectivity, reshaping communication, and fostering the rise of digital culture. Personal computing democratized access to technology, changing how people work, learn, and interact. These technologies also catalyzed the emergence of new economic models, such as e-commerce and the gig economy (which is a labor market characterized by the prevalence of short-term contracts or freelance work).

 

AI's Growing Influence: The trajectory of AI's impact on society has evolved from basic automation to complex decision-making. Initially, AI was used for routine tasks, like calculations and data processing (Kaplan, 2016, p. 144). Over time, advancements in machine learning and neural networks have enabled AI to tackle more sophisticated tasks, including pattern recognition, natural language processing, and predictive analytics. This evolution has led to AI systems that can make decisions in complex environments, influencing fields like healthcare, finance, and transportation.

 

Technological Determinism and Generative AI

Through the lens of technological determinism, generative AI can be seen as significantly shaping societal norms and behaviors, as it is influencing various facets of life.

 

1. Creativity and Art: The emergence of AI in creative fields is redefining the concept of creativity, traditionally seen as a uniquely human trait. AI-generated art and literature challenge our understanding of creativity and originality (Boden, 2016, pp. 57-59). This technological advancement is not only creating new forms of art but is also influencing how people perceive and interact with artistic works.

 

2. Work and Employment: Generative AI is transforming the workplace. It automates tasks, thus shifting the nature of jobs and required skills (Brynjolfsson & McAfee, 2014). This shift could lead to job displacement in certain sectors while creating new opportunities in others, fundamentally altering employment landscapes.

 

3. Ethics and Society: The capabilities of generative AI bring forth ethical dilemmas, especially concerning data privacy, the authenticity of information, and the potential for misuse in creating deepfakes. These concerns necessitate a reevaluation of ethical frameworks and legal standards to keep pace with technological advancements.

 

4. Media and Communication: In media, generative AI's ability to produce realistic content is transforming how information is created and consumed. This raises concerns about the authenticity of information and the potential for spreading misinformation (Tufekci, 2015). As AI becomes more involved in content creation, it also influences the way narratives are shaped and disseminated.

 

Current Generative AI Applications

Generative AI is increasingly being used in journalism, although its impact and the extent to which it can replace human authors are subjects of ongoing debate and exploration.

Several news organizations are experimenting with AI for various aspects of journalism. The Associated Press, for instance, has created a detailed module with specific guidelines for using AI, employing it for tasks like compiling digests of stories for newsletters and creating short news stories from sports scores or corporate earning reports (Bauder, 2023). AP stated that any item produced by AI must be “carefully vetted - just like material from any other source, and that a photo, video, or audio segment generated by AI should not be used unless that segment is the subject of a story itself” (Hurst, 2023).

The Guardian (2023), on the other hand, has outlined its approach to generative AI, focusing on using the technology to assist journalists in managing large data sets, with strict human oversight and a senior editor's permission required for any editorial use of AI. Similarly, local newsrooms are exploring AI to publish a high volume of local stories on topics such as weather, fuel prices, and traffic conditions, as seen with News Corp Australia's production of 3,000 articles a week using generative AI.

 

However, the role of AI in journalism is not without challenges. Ethical considerations and the potential for factual inaccuracies are major concerns. For instance, tech outlet CNET faced criticism for publishing AI-generated content without clear disclosure, leading to an update in their processes for greater transparency. Not only did CNET publish AI-generated material, but it also created “articles generated by artificial intelligence, on topics such as personal finance, that proved to be riddled with errors” (Harrington, 2023).

 

Ethical guidelines suggest that AI-generated content should be clearly disclosed to audiences and not presented as human-written. Additionally, there are challenges related to the accuracy of information, especially in breaking news reporting, as AI models often struggle with generating accurate and factual information regarding current events or real-time data.

 

Overall, while generative AI presents opportunities for enhancing productivity in journalism, it also requires careful consideration of ethical, human, and editorial implications. The technology is viewed not as a replacement for human journalists but as a tool to augment their capabilities, allowing them to focus on tasks that require human judgment and creativity. Therefore, the complete elimination of authors by AI in journalism is not currently foreseeable, given the technology's limitations and the value placed on human insight and analysis in the field.

 

Societal and Ethical Implications

The implications of AI-generated content on concepts like authorship, intellectual property, and truth in media are complex and multifaceted.

In terms of authorship and intellectual property, the rise of AI-generated content has led to significant legal challenges. A fundamental issue is determining the actual creator of AI-generated works and, consequently, who owns the copyright. Under U.S. copyright law, generally, the creator of the content owns the copyright, but this becomes complicated when an AI algorithm creates the work. For instance, in the case of Thaler v. Perlmutter et al., the court upheld the United States Copyright Office’s decision that human authorship is a prerequisite for valid copyright protection. This decision underscores the importance of human creativity in copyright law but leaves unresolved how to handle content created from both AI and human input (Clarida and Kjellberg, 2023)

There's also the concern of potential plagiarism or copyright infringement with AI-generated content. “The instances of academic plagiarism have escalated in educational settings, as it has been identified in various student work, encompassing reports, assignments, projects, and beyond” (Elkhatat, Elsaid and Almeer, 2023). AI writing assistants, while designed to generate original content, could inadvertently produce work substantially similar to existing material, potentially leading to accusations of plagiarism or copyright infringement. In such cases, "I didn't know" is not a viable defense, as most forms of infringement are strict liability torts. This highlights the inherent risk in using AI for content creation, as users often cannot verify the source of information or ensure content originality.

Regarding the truth factor in the media, the use of AI in journalism raises ethical and factual accuracy concerns. While AI can assist in synthesizing information and informing reporting, its current capabilities lack originality, analytical skills, and a developed voice, essential for quality journalism. Another factory that should be stressed is that AI “is a ‘language machine…not a truth machine’, so the human factor is still a vital element in producing journalism” (Hurst, 2023). Moreover, AI models often struggle with generating accurate and factual information, particularly in real-time or current events, posing a challenge for breaking news reporting. Thus, while AI has a role in journalism, it cannot solely be relied upon, especially for complex and nuanced reporting.

In conclusion, while AI-generated content offers many opportunities for innovation and efficiency, it also brings significant challenges in authorship, intellectual property, and maintaining the integrity of information. These challenges necessitate careful consideration and adaptation of legal and ethical frameworks in the digital age.

 

Case Study: The Associated Press

A specific example of generative AI impacting journalism is the use of AI-driven tools in the newsroom of the Associated Press (AP). The AP has integrated AI into its journalistic processes, primarily for automating the creation of straightforward news reports, especially in areas like sports and finance.

 

The AP uses AI to automatically generate news stories from structured data. This began with their use of a tool called Wordsmith, developed by Automated Insights, to produce news stories on corporate earnings reports. By inputting data into Wordsmith, AP was able to automate the creation of earnings reports articles, a task that was previously time-consuming for human reporters (Lewis-Kraus, 2016).

 

This automation significantly increased productivity. Before implementing AI, AP reporters wrote about 300 earnings reports stories per quarter. After the adoption of AI, this number increased to over 3,000, demonstrating a tenfold increase in output without sacrificing accuracy (Philips, 2013). Moreover, this automation freed journalists to focus on more complex, investigative stories where human insight and analysis are irreplaceable.

 

However, the implementation of AI in journalism also raises concerns regarding job displacement and the potential for errors in automated content. While AI has enhanced efficiency in news production, it has also led to debates about the evolving role of journalists in an increasingly automated news environment (Graefe, 2016).

 

The AP's approach to AI in journalism reflects a broader trend in the industry: leveraging AI for routine, data-heavy tasks, while retaining human journalists for more nuanced and analytical work. This strategy underscores the complementary role of AI in journalism, augmenting human capabilities rather than replacing them entirely.

 

Benefits:

- Increased Productivity: The use of AI has allowed the AP to increase the volume of content produced. For instance, their earnings reports coverage expanded from 300 to over 3,000 articles per quarter after implementing AI (Philips, 2013).

- Resource Allocation: By automating routine reports, AI frees up journalistic resources, allowing human reporters to dedicate more time to in-depth, qualitative reporting (Graefe, 2016).

 

Challenges:

- Accuracy and Reliability: While AI improves efficiency, there are concerns in regards to accuracy of the generated content, especially in complex or nuanced reporting scenarios.

- Ethical and Employment Concerns: The integration of AI in journalism also raises ethical questions about transparency and the potential for job displacement in the industry (Lewis-Kraus, 2016).

 

Broader Impact:

- Impacts on Journalism: AI is transforming the journalism industry by changing how news is produced and consumed. It encourages a shift towards more data-driven journalism and may change the skill sets required for future journalists.

- Societal Implications: The widespread use of AI in media can influence public perception and understanding of news, underscoring the need for clear guidelines and ethical standards in AI-generated content (Graefe, 2016).

 

Conclusion:

The adoption of generative AI by organizations like the AP highlights both the potential and challenges of this technology in journalism. While it enhances efficiency and allows journalists to focus on more complex and high-quality reporting, it also brings up questions about the future of the human factor, the accuracy information, and ethical considerations in media. As this technology evolves, its integration into journalism will likely continue to influence both the industry and societal perceptions of news and information.

When considering the use of generative AI in journalism, exemplified by the Associated Press and its adoption of automated news writing/generation, it should be stated that it can be analyzed through the lens of technological determinism because it confirms that fact that technological advancements play a primary role in shaping societal structures, cultural norms, and human behavior (McLuhan, 1964, pp. 7-8). In the context of journalism, the integration of AI technologies aligns with this theory in several ways:

- Shaping News Production: The adoption of AI for routine news generation signifies a complete modification of the journalistic processes, driven by technology. The increased efficiency and capacity for producing large volumes of content demonstrate how technology can redefine industry practices in general. (Philips, 2013).

- Influencing Journalistic Roles: As AI takes over more routine and data-driven tasks, the role of journalists in the field evolves. This aligns with technological determinism, where technology influences human roles and skills required in a profession (Graefe, 2016).

- Impacting News Consumption: The way audiences consume news can also be influenced by the presence of AI-generated content, potentially leading to changes in how people interact with and perceive news media. This is a direct implication of technological change influencing societal behavior, a core concept of technological determinism.

- Ethical and Societal Considerations: The ethical concerns and the potential for misinformation with AI in journalism highlight the broader societal impacts of technology. These implications reflect technological determinism's assertion that technology not only changes practices but also raises new ethical and societal questions (Lewis-Kraus, 2016).

 

In summary, the application of generative AI in journalism and its subsequent effects on the industry and society exemplify the principles of technological determinism. The technology is not merely a tool but a transformative force that reshapes industry norms, professional roles, and societal interactions with news media.

 

 

Bibliography:

 

Ahmed M. Elkhatat, Khaled Elsaid & Saeed Almeer (2023) Evaluating the efficacy of AI content detection tools in differentiating between human and AI-generated text. Available at: https://edintegrity.biomedcentral.com/articles/10.1007/s40979-023-00140-5/ (Accessed: 17 November 2023).

 

Bauder, D. (2023) AP, other news organizations develop standards for use of artificial intelligence in newsrooms. Available at: https://apnews.com/article/artificial-intelligence-guidelines-ap-news-532b417395df6a9e2aed57fd63ad416a/ (Accessed: 28 November 2023).

 

Boden, M. (2016) AI: Its nature and future. Oxford, the UK: Oxford University Press.

 

Brynjolfsson, E. and McAfee, A. (2016) The Second Machine [Ebook]. New York: W. W. North & Company.

 

Castells, M. (2000). The Rise of the Network Society. Blackwell Publishers.

 

Clarida, R. and Kjellberg, T. (2023) ‘Thaler v. Perlmutter’: AI Output is Not Copyrightable. Available at: https://www.law.com/newyorklawjournal/2023/09/14/thaler-v-perlmutter-ai-output-is-not-copyrightable/ (Accessed: 22 November 2023).

 

Graefe, A. (2016) Guide to automated Journalism, Tow Center for Digital Journalism, New York: Tow Foundation.  Available at: https://academiccommons.columbia.edu/doi/10.7916/D8QZ2P7C/download/  (Accessed: 14 November 2023).

 

Harrington, C. (2023) CNET Published AI-Generated Stories. Then Its Staff Pushed Back. Available at: https://www.wired.com/story/cnet-published-ai-generated-stories-then-its-staff-pushed-back/ (Accessed: 20 November 2023).

 

Hurst, L. (2023) Robot reporters? Here’s how news organisations are using AI in journalism. Available at: https://www.euronews.com/next/2023/08/24/robot-reporters-heres-how-news-organisations-are-using-ai-in-journalism/ (Accessed: 20 November 2023).

 

Kaplan, J. (2016). Artificial Intelligence: What Everyone Needs to Know. Oxford, UK: Oxford University Press.

 

Lewis-Kraus, G. (2016). The Great A.I. Awakening. Available at: https://www.nytimes.com/2016/12/14/magazine/the-great-ai-awakening.html/ (Accessed: 15 November 2023).

 

McLuhan, M. (1964). Understanding Media: The Extensions of Man. New York: McGraw-Hill (PP. 7-8).

 

Mike Loukides (2023) What Are ChatGPT and Its Friends? California, USA: O’Reilly Media.

 

Philips, M. (2013) How the Robots Lost: High-Frequency Trading's Rise and Fall. Available at: https://www.bloomberg.com/news/articles/2013-06-06/how-the-robots-lost-high-frequency-tradings-rise-and-fall/ (Accessed: 19 November 2023).

 

Tufekci, Z. (2017). Twitter and Tear Gas: The Power and Fragility of Networked Protest. Connecticut, USA: Yale University Press.

 

Turing, A. M. (1950). ‘Computing Machinery and Intelligence.’ Mind, 59(236), 433-460.

 

Viner, K. and Bateson, A. (2023) The Guardian’s approach to Generative AI. Available at: https://www.theguardian.com/help/insideguardian/2023/jun/16/the-guardians-approach-to-generative-ai/ (Accessed: 24 November 2023).