By Lorena Mendez
@redlineproject
AI disclosure: Perplexity was used to outline a segment of this paper and for research. DALL-E was used for image generation, and Sora for video generation. Notebook LM and Nota were used to summarize and provide an FAQ at the end.
—
Summary: Artificial intelligence in newsrooms raises bias and transparency concerns. While AI can help journalists with data analysis, it can also lead to misinformation and damage credibility. The Associated Press has partnered with OpenAI to license its text archive, which could further reinforce existing biases. Journalists must disclose their use of AI and explain why they use it to maintain trust with their audience. The future of journalism depends on a balanced approach that embraces the potential of AI while ensuring ethical standards and transparency.
—
About 90% of Americans said they have at least heard a little about artificial intelligence (AI), according to a November 2023 study conducted by the Pew Research Center. From that study, we learned that many more Americans were feeling increasingly cautious about the growing presence of AI in their lives. Interestingly, only 30% of U.S. adults could pinpoint specific uses of AI, which signals the significant gaps in understanding the fundamental concepts of AI and its capabilities.
AI is still relatively new to the average person in America, though it has existed since the 1950s. In a way, American audiences need to be convinced that if journalists use this tool in their newsrooms, they are doing so ethically and being extremely transparent in its use.
AI is expanding in unprecedented ways, and it has set off many alarm bells for people in different industries who are afraid of the unknown. The public has also called to question AI’s capabilities with its rapid rise, how much it will impact their lives, and how they view and get information. Given all the rapid change, the impact of the investments made by AI companies on newsrooms is called into question.
We know how crucial it is for journalists to report the news ethically and to get the facts right every single time, but how can we be sure that will be the case with the growing amount of AI embedded in newsrooms? This paper will explore how investments made in newsrooms can further certain biases and damage the relationships between the audience and the journalist. If AI were to be the sole author of news stories today, how would it fairly translate the news and be able to prioritize particular perspectives? This on its own is dangerous. How would it translate the news and be able to prioritize specific perspectives and, most importantly, combat misinformation?
The rapid change in AI is shaping journalism; we have seen what an essential tool AI has been in at least the last decade. Journalists are known for sorting through pages and pages of corporate earnings or court records and various reports, making data analysis exhausting, so why not use a tool to help them do this process more efficiently? The public can see this as misleading and would demand full transparency for its use. Since AI is such a user-friendly tool, it’s easy for readers to feel like information and data are easily manipulated, creating story inaccuracies. This is dangerous because it can cause news organizations to lose credibility and trust with their audience. Journalists are up against a lot; they must deliver the news accurately, so when one of their AI tools produces misinformation, there needs to be a procedure to ensure that information is corrected promptly.
—
PODCAST: Listen to a short AI-generated podcast about the growing presence of AI in newsrooms and how to balance public trust and bias
—
AI Investments
In recent years, there have been growing investments from AI companies in newsrooms, most notably a deal struck between OpenAI and Axios. The interesting aspect of this investment is that it is concentrated only in local newsrooms, which can have adverse, lasting effects. Knowing that AI is being used in more concentrated levels leaves room for existing biases and could reinforce preexisting stereotypes in these local communities. This is already problematic since specific demographics find it difficult to accept different viewpoints and narratives.
In a 2025 study by Dr. Gregory Gondwe from California State University-San Bernandino, he argues that AI can be a double-edged sword. AI can help improve journalistic practices and present opportunities for newsrooms to grow, but it can also present problematic narratives. We must remember that the role of journalists is to uncover truths and dig deep into stories so they have a human element to them. Where does all that go if newsrooms integrate AI fully into their journalistic work?
Gondwe argues that it is the job of the journalist to be able to point out these harmful biases and further train their AI system to weed out that information and have leverage over AI to improve their work. But again, what I call to question is the biases that exist in the journalist who is monitoring these systems. This sounds like a process where there needs to be constant checks and balances every step of the way, where AI audits the journalist and vice versa. Gondwe states that “the future of journalism will depend on a balanced approach that embraces the potential of AI.” We must make AI human-centric and apply The Associated Press (AP) “80/20 rule,” where 80% can be automated, and 20% requires human interaction (i.e., editing, fact-checking, editing code).
Many newsrooms and media outlets have been using AI to report the news, but it was done in ways where their creditability wasn’t questioned. Some big names in the news world, such as The Washington Post, have been using generative AI since 2016 when they used their generative AI tool Heliograf to help cover the 2016 Olympic Games in Rio de Janeiro. The Post assured the public it was being used in more data-centric news bits. While not raising so many red flags, the publication defended its choice by making the case that journalists should focus their energy on more “interesting and complex work.”
Yet, there are no guarantees that AI will remain just a tool to journalists. This is causing people to feel annoyed and anxious with the use of AI in the news, according to a focus group with Poynter and the University of Minnesota. Participants from the focus group believe that journalists using AI is almost like flat-out cheating, and large segments of the public feel these same sentiments. On the other hand, others don’t know how to feel about these AI developments and what they mean to society.
Issues with Biases
The Associated Press also has signed a deal with ChatGPT/OpenAI. The agreement includes that OpenAI will license a portion of the AP’s text archive, which, on the surface, sounds scary. This means that OpenAI will be allowed to further train on the Associated Press text archives to train its language model for its AI product ChatGPT. According to details released on the partnership, the AI company will have access to archives dating back to 1985. Now, this can be problematic on many levels since the political and racial landscape of America was quite like it is today. There was heightened racial tension, and the War on Drugs was still going on. These datasets will be fed into a system that can form many biases with that information alone. If these are just some datasets that ChatGPT will train on, how will this be a helpful tool that won’t cause harm when used in newsrooms?
It is hard to tell what biases will be drawn from the Associated Press, but there is plenty to go on. Since the AP is the only news organization with this type of partnership, what does that read to the public? They have done a good job trying to reassure the public that they are not using AI to write stories, but instead, they are using ChatGPT as a behind-the-scenes tool that will help journalists sort through data and generate reports and simply assist journalists with research. It all sounds harmless on the surface, but this is something that the AP has already been doing for nearly a decade, where they were using generative AI to sort through corporate earnings and recap sporting events. Why is there a big commitment now?
Obviously, we are moving into a time when some of this technology will be impossible to avoid. Still, there needs to be more transparency between news organizations and the public since they are becoming a lot more aware of the power of AI. There is no question that digitization has affected journalists, but this isn’t the first time they have had to adapt, and they will continue to do so. It is fair to say that their reputation is on the line while working alongside AI.
Moreover, the speed at which AI can work is an advantage to journalists and the companies they work for, but are the drawbacks worth it? The most significant disadvantages can be the spread of misinformation and the chance that journalist ethics can be violated. This risk is too substantial to take in today’s media landscape, as the audience already has difficulty trusting the news they watch. According to Trusting News in its Trust Kits on AI, journalists must disclose that they are using AI in their work, then take it further and explain why AI was used. Doing this creates a more trusting relationship, makes people less apprehensive about using AI in news stories, and, most importantly, keeps their credibility intact.
Trusting News studies ways newsrooms can better understand and build trust with audiences. Trusting News found that 93.8% of news consumers surveyed by Trusting News cohort said they want the use of AI to be disclosed in journalism, and they want details about that use. In other words, the ball is in the newsrooms’ court if they want to keep the trust of their audience, they must follow some very crucial steps. Furthermore, the cohort also found that the participants found it necessary for humans to take part in the editing process by educating the public on the use of AI and the newsroom to provide a public service, creating a partnership and empowering the community, and inviting them to actively participate in shaping the future of journalism.
Conclusion
As more Americans and people worldwide become more familiar with AI and its capabilities, we must approach the information we see more cautiously. There will continue to be mistrust and even more confusion surrounding AI until concrete ethical standards are established to ensure that people who intend to use it maliciously can be stopped. AI will continue to be used in newsrooms, not just in the U.S. but worldwide, making it much more challenging to get a handle on misinformation. The investments made into newsrooms will only increase, but how those investments are made, and their impact on journalism and readership remains to be seen.
The risks and concerns of the narratives that can come from AI are only gaining more fear from the public. Especially the biases that already may be present in datasets that are being used to train AI. The datasets that are being used can train on harmful information, especially in the small newsrooms that currently have partnerships with AI companies. Again, as I have mentioned throughout this paper, AI is not being used to write stories; it is simply being labeled as a tool that assists journalists in research and other mundane tasks.
Yes, journalism is rapidly changing, and many controversies will follow with the adaptation of AI. Still, the lack of human judgment will ultimately decide if AI and journalism can work harmoniously. Just as many view AI as a tool, I hope people understand that AI can help us in our everyday lives. I’m convinced that we as a society need to take the role of using AI seriously, just like anything else, and use it responsibly. I am too fearful of what AI will look like in the future, the cultural shifts, and the possible racial divides it can create, but more than anything, I’m excited to see how far AI will take us as a society.
—-
Frequently Asked Questions about AI in Journalism
1. How aware is the American public about artificial intelligence (AI), and what are their general sentiments towards its increasing presence?
A Pew Research Center study in November 2023 indicated that approximately 90% of Americans have heard at least a little about AI. However, many harbor increasing caution regarding its growing presence in their lives. Notably, only about 30% of U.S. adults could identify specific uses of AI, highlighting a significant lack of understanding about its fundamental concepts and capabilities. This suggests that while awareness is high, comprehension and comfort levels are considerably lower, leading to apprehension about its integration into various sectors, including journalism.
2. What are some perceived benefits of using AI in newsrooms?
AI is potentially valuable for journalists in several ways. It can enhance efficiency by helping to sort through large volumes of data such as corporate earnings, court records, and various reports, making data analysis less time-consuming. This allows journalists to dedicate more time to more complex and in-depth aspects of their work. Additionally, AI can assist with research and the generation of reports, acting as a behind-the-scenes tool to support journalistic endeavors rather than replace human writers entirely. Some news organizations have already used AI for data-centric tasks like covering sporting events or the Olympics.
3. What are the primary concerns and potential risks associated with the growing use of AI in journalism?
Several significant concerns exist regarding the integration of AI into journalism. One major worry is the potential for AI to introduce or amplify biases present in the datasets it is trained on, leading to skewed or unfair reporting, particularly in sensitive areas like local news. There are also fears that AI could be used to spread misinformation or create inaccuracies in news stories due to data manipulation or flawed algorithms, which could severely damage the credibility and trust of news organizations. Furthermore, the public expresses anxiety that the increasing use of AI might erode the human element and ethical considerations central to quality journalism.
4. How do investments by AI companies in newsrooms, such as the deal between OpenAI and Axios, raise concerns?
The increasing investments from AI companies into newsrooms, mainly when concentrated in specific areas like local news, raise concerns about potential adverse effects. Focusing AI tools in local newsrooms could exacerbate existing biases and reinforce stereotypes within those communities, especially given that some demographics already struggle with accepting diverse viewpoints. The partnership between OpenAI and Axios, which focuses on local news, exemplifies this concern. Additionally, licensing news archives, such as the Associated Press’s text archive to OpenAI, raises worries about the AI models being trained on potentially biased historical data, which could then influence the AI’s output when used in news production.
5. How is the ethical use of AI in journalism and the maintenance of audience trust being addressed?
Recognizing the public’s apprehension, there is a growing emphasis on the ethical use of AI in journalism and the need for transparency to maintain audience trust. News organizations are urged to disclose when and why AI is used in their work, providing details about its application. This open communication is crucial for building a more trusting relationship with the audience and mitigating fears of manipulation or inaccuracy. Furthermore, guidelines like the Associated Press’s “80/20 rule,” which suggests a balance between AI automation and human oversight (editing, fact-checking), are being promoted to ensure human judgment remains a critical component of the news production process.
6. What role do journalists and news organizations have in ensuring the responsible implementation of AI in their work?
Journalists and news organizations are responsible for ensuring the responsible implementation of AI. This includes actively monitoring AI systems for biases and inaccuracies and critically evaluating their output. It also involves continuously training AI systems to mitigate harmful biases and exercising human leverage to improve their work. Implementing constant checks and balances, where AI and journalists audit each other, is suggested to maintain accuracy and ethical standards. Ultimately, newsrooms are encouraged to adopt a balanced approach that leverages the potential of AI while upholding the core values of journalism, including accuracy, fairness, and a human-centric perspective.
7. What are the public’s expectations regarding the disclosure and oversight of AI in news production?
Surveys indicate that news consumers overwhelmingly desire transparency regarding the use of AI in news production. A significant majority (93.8% according to one survey) want news organizations to disclose when AI is used and provide details about its specific application. Additionally, the public believes that human involvement in the editing process is essential. They also see a role for newsrooms in educating the public about AI’s use, providing a public service, fostering community engagement, and inviting active participation in shaping the future of journalism.
8. What does the future of journalism likely look like with the increasing integration of AI, and what are the key challenges and opportunities?
The future of journalism with increasing AI integration will likely involve a significant shift in how news is gathered, processed, and presented. AI offers opportunities for enhanced efficiency in data analysis and research, allowing journalists to focus on more complex and nuanced storytelling. However, key challenges remain, including the risk of bias amplification, misinformation spread, and public trust erosion if AI is not implemented ethically and transparently. Ensuring a harmonious relationship between AI and human judgment will be crucial. The future success of AI in journalism will depend on establishing concrete ethical standards, maintaining human oversight, and fostering open communication with the public to address their concerns and harness the benefits of this rapidly evolving technology responsibly.