Moor Studio/Getty Images Plus
Artificial intelligence’s (AI) explosion in popularity has spanned nearly every industry, acting as a catalyst for rapid transformation across the makeup of many sectors. The media and journalism world is no different, adopting AI to increase efficiency and convert large sums of information into digestible outputs for the general public. Utilized to expedite transcriptions, facilitate content production and drafting, and assess audience analytics, AI has become a powerful tool for many journalists. However, the negative implications of AI implementation into journalism are twofold: replacing human journalists with machines, and compromising the integrity of journalism as a whole. Absent oversight or guiding standards, these developments could undermine the five values of ethical journalism—accuracy, independence, impartiality, humanity, and accountability—destabilizing the foundation of free and open media.
In terms of replacement, the field of journalism is experiencing a period of mass layoffs. Whether these layoffs are a result of AI’s growing popularity in the industry, or conversely, AI is being utilized as a means to lessen the load on short-staffed outlets, there is an undeniable relationship between the two. While some argue that AI is simply a supplemental tool in journalism, not a replacement mechanism, the phenomenon of automation bias across many various manifestations of AI remains problematic. The human tendency to over-rely on automation can completely overtake human decision-making for the sake of expediency and ease. For example, younger generations are losing the ability to read physical maps in favor of putting their full faith in navigation apps. This blind trust can lead to disastrous situations so common that they’ve earned their own moniker: “death by GPS.” In journalism, automation bias can mean reporters spend less time verifying AI-generated content, inclined to trust it at face value despite generative AI (GenAI) needing significant human oversight due to its experimental nature.
Additionally, layoffs in journalism disproportionately impact marginalized groups, specifically people of color and women. This issue of declining diversity in journalism mirrors the recent pushback against Diversity, Equity, and Inclusion (DEI) initiatives spearheaded by the Oval Office. The devaluation of marginalized voices is problematic in any context, but in the media field particularly, a reduction in perspectives can create an environment conducive for harmful misinformation and inaccurate representations. Replacing journalists with AI exacerbates the potential for extremely biased reporting, due to the fact that GenAI models are commonly known to amplify both racial and gender-based stereotypes. Without someone in the room to add their lived experience and nuance to the conversation, journalists may unknowingly perpetuate negative stereotypes or greenlight AI-generated content that does.
Journalists are already fighting an uphill battle against AI-generated misinformation. Falsely generated AI news and deepfakes have made it increasingly difficult for journalists to verify facts in their reporting. These technologies have the power to sway public opinion and quickly spread false information during crucial times, such as crises and elections. AI’s use on both ends, for content creation and content verification, manufactures a cyclical media landscape dependent on AI. This becomes an epidemic of “platformization” of newsrooms, due to tech giants like Google and Microsoft selling newsroom AI products that can render publications completely dependent on Big Tech for their journalistic processes. Preserving the integrity of unbiased and truth-based reporting is becoming more and more crucial as social media platforms are overrun with unregulated misinformation.
As previously mentioned, AI-produced outputs necessitate human oversight to catch any errors born from the nature of models trained on the Internet; troves of both factual and fake information live on the Internet, which ChatGPT and other GenAI models indiscriminately draw upon to craft their responses. With this comes an increased risk for AI to plagiarize sources without accreditation, unbeknownst to the journalist using the output for their own publications. GenAI is also known to “hallucinate” by creating and dispensing baseless information as fact; ChatGPT has even fabricated entire articles, and then tacked on the names of real reporters as the authors. When adopted into media environments, GenAI’s implementation muddies the world of credit attribution and factual integrity, while simultaneously pressuring journalists to prioritize speed over accuracy. Accelerating the processes of journalism with AI leads to higher competition to break stories first, which can reduce time spent on necessary fact-checking and verification.
The most recent developments regarding AI and journalism come from OpenAI; while already enmeshed with 19 popular news publishers, OpenAI is now moving to directly fund local Axios newsrooms enabled by OpenAI products. The partnership’s ultimate vision is an AI “super-system” that ascends beyond the one company, and would quality-control editing, create visuals for articles, and control distribution of articles.
It seems this super-system is already materializing in some respect, with President Trump’s endorsement and partnership in the $500 billion AI infrastructure venture with a company called the Stargate Project. The partnership extends across borders, consisting of OpenAI, Oracle, Japan's Softbank, and the United Arab Emirate’s (UAE) MGX. This would fund massive AI data centers in the US, and supposedly generate hundreds of thousands of American jobs. However, the origin of the $500 billion is up for debate, with Elon Musk commenting that “they don’t actually have the money.” Alternatively, one source claims that the bulk of funding is coming from the technology arm of the UAE’s sovereign wealth fund. Such significant foreign funding in our media and content-producing sphere is cause for concern, especially when considering multiple countries’ attempts to meddle in US affairs in the past.
President Donald Trump’s policy stance on AI remains consistent with his enthusiasm towards the Stargate Project, seeing as he just signed an executive order rescinding former President Joe Biden’s 2023 executive order that sought to establish guardrails and standards for AI usage and development. Biden’s extensive executive order touched on many aspects impacted by AI, requiring transparency from prominent AI developers, standards of safety and security created by the National Institute of Standards and Technology, as well as stipulations pertaining to privacy, consumer protections, and civil rights. Trump’s executive order “calls for departments and agencies to revise or rescind all policies, directives, regulations, orders, and other actions taken under the Biden AI order that are inconsistent with enhancing America’s leadership in AI.” In other words, anything inhibiting or hindering the profit and expansion of the AI industry in the US is to be effectively eliminated.
President Trump’s coziness with Big Tech presents another alarming layer to this issue. Trump is already in cahoots with Meta, Tiktok, and X, so the link between Trump, OpenAI, and newsrooms like Axios becomes particularly troubling. With the end of fact-checking across Meta platforms, and the rapid dissemination of misinformation on social media in general, the importance of reputable journalistic reporting is more essential now than ever.
The implementation of AI into journalism must be done with intentional and careful considerations of the advantages and disadvantages of the tool, as well as clear guidelines for use and credit attribution. Transparency in how, when, and why AI is utilized must become the standard. Otherwise, we risk devolving into a period where reputable reporting is nonexistent or highly inaccessible. At a time of such heightened political tensions and ever-evolving current events, protecting the integrity of journalism must be a priority.