In the latest edition of Eye on AI, several key topics are highlighted, including the impact of AI on the news media industry, an order from former President Trump to refocus U.S. AI safety efforts on ideological bias, and the growing interest in distributed training. The discussion also touches on concerns about increasingly powerful AI contributing to the rise of totalitarianism.
AI is emerging as a potentially disruptive force across various business sectors, with the news industry facing particularly existential threats. The role of the free press is fundamental to democracy, emphasizing the significance of these challenges. There are parallels between how news executives are addressing these AI challenges and opportunities and the strategies being adopted by leaders in other sectors.
Recently, a conference titled “AI & News: Charting the Course” was held at Reuters’ headquarters in London. This event, hosted by the Aspen Institute, included top executives from UK and European news organizations and was conducted under Chatham House Rules, meaning specific individuals’ statements could not be attributed. However, the discussions centered on utilizing AI to enhance internal efficiency, such as using AI for writing search engine-optimized headlines and translating content to reach broader audiences, while maintaining human oversight for accuracy.
One editor shared the use of AI to automate the production of short articles from press releases, thereby freeing up journalists for original reporting but ensuring human editors oversee quality control. Journalists also utilize AI for summarizing documents and analyzing large datasets, which supports investigative journalism otherwise difficult without such tools. These applications primarily improve existing workflows.
There was debate at the conference regarding whether news organizations should adopt a bottom-up or top-down approach to AI integration. A bottom-up strategy empowers journalists and editors with generative AI tools for data analysis and creating AI-powered widgets to assist their work. While this democratizes access, it poses challenges in terms of compliance with ethical and legal standards and managing technical debt. Conversely, a top-down approach may ensure a more organized strategy but could stifle innovation. Many participants advocated for a balanced approach without reaching a clear consensus.
Caution surrounds the adoption of audience-facing AI tools by news organizations. Some have begun using AI to generate bullet-point summaries and deploy chatbots to answer questions within specific coverage areas, labeling such initiatives as “experiments” to signal potential inaccuracies. Concerns persist that AI-generated content might undermine trust in journalistic accuracy, with brand reputation and business models reliant on this trust.
The hesitation to embrace AI fully poses a significant risk. With technology companies increasingly employing AI to summarize news and provide more interactive experiences, news organizations risk losing direct audience connections—a phenomenon known as “disintermediation.” Many news executives fear this loss, as they rely on Google searches and other platforms to reach their audiences. Studies reveal notably lower clickthrough rates for AI-generated searches compared to traditional methods.
So far, news organizations have responded to these threats by pursuing legal actions and developing partnerships. Legal cases against companies like OpenAI and Perplexity allege copyright violations, while partnerships involve significant licensing deals. These arrangements offer revenue from content already being used by tech companies and establish collaborative relationships to harness AI expertise. However, reliance on AI licensing revenue poses long-term risks reminiscent of past over-dependence on platforms like Facebook.
The conference noted the necessity for news organizations to build direct audience relationships independent of AI intermediaries. One expert criticized the industry’s tendency toward incremental rather than transformative adaptation, drawing an analogy to a process anchored by an “AI-powered Ferrari” at both ends, interspersed with a “horse cart” in the middle. There is advocacy for moving beyond traditional article-centric approaches, instead utilizing generative AI to transform various media formats into diverse outputs tailored to audience preferences.
Some industry advisors suggested learning from other sectors and even fostering startups to explore new business models for the AI era. The stakes are high, as AI poses existential challenges and unprecedented opportunities for journalism, offering a chance to reconnect with audiences if leaders are willing to reimagine the role of news in the AI age.
In other AI news, Jeremy Kahn, from Fortune, notes a couple of corrections to previous editions regarding the headquarters of Trustpilot, confirmed to be Denmark, and the correct name of the Chinese startup behind the viral AI model Manus—Butterfly Effect. This article originally appeared on Fortune.com.
