On Monday, September 9, guest writer Pramit Ghatak attended Columbia University’s School of Journalism’s panel discussion titled “AI & Local News: Challenges and Opportunities for Journalism in NY.” The panel focused on the increasing role of Artificial Intelligence (AI) in the news industry.

The ongoing discussion about AI and journalism is necessary to navigate the integration of non-human technology into a field that increasingly values human connection and personal engagement between news staff and audiences. Columbia’s School of Journalism panel focused on the growing need to balance the use of AI with maintaining meaningful relationships between both. The discussion was moderated by Professor Juan Manuel Benitez, Columbia’s Professor of Local Journalism, and featured panelists Kristen Gonzalez, New York State Senator; Zachary Richner, founder of Empire State News Coalition; Elinor Tatum, publisher of the Amsterdam News; Marjorie Velazquez, Vice President of Policy at Tech: NYC; and Mark Hansen, Professor of Journalism and Innovation at Columbia.

The conversation opened with a discussion about the traditional perceptions of AI in the media. Velazquez noted that many people view AI through the lens of pop culture depictions, such as the cyborg villain in The Terminator (1984). However, she emphasized that AI models can be used in ways that enhance human interaction, asking, “How do we make it better for journalists? How can we get technology to help, not hurt?”

This perspective, that AI can assist journalistic work, is complicated by the growing pressure to regulate its use due to concerns about AI plagiarizing and exploiting journalists’ work. Hansen explained that local newsrooms’ only line of defense against automated systems harvesting and plagiarizing their content is a simple file called robots.txt, which essentially asks bots not to steal content. Surprisingly, asking robots designed to steal content to please not steal hasn’t been very effective.

Despite the challenges of plagiarism, the panelists recognized AI’s potential to spotlight historically marginalized voices. Tatum highlighted AI-driven search engines like Lattimore. This software was designed to serve as a competitor to ChatGPT, which draws on African American journalism archives as its primary sources. The search engine allows users to access a historical perspective different from the predominantly Eurocentric bias that search engines typically offer. As Tatum puts it, “Whoever’s distilling information for you is the perspective you’re going to get.”

While AI has the potential to both harm and help journalists, all the panelists agreed on the need for more legislation and ethical guidelines surrounding its use. Senator Gonzalez, chair of the New York Senate’s Internet and Technology Committee, discussed her work on the New York Fair Act, legislation aimed at regulating the use of celebrity deepfakes by holding journalists accountable for their creation. She also acknowledged the moral complexities of whether the blame lies with the journalists using AI or the companies developing the technology. Gonzalez also sponsors the LOADinG Act, which would regulate the use of AI in political campaigns and ads.

Hansen stressed that the ethical responsibility of using AI rests largely with journalists themselves. He emphasized that journalists should be trained to develop AI prompts to assess bias, not just to generate content, but to refine their understanding of the “right” questions to ask for better answers, and thus, better articles. AI’s role in journalism is here to stay. New York State and other private and public sources have invested $400 million into AI literacy and education at universities like Columbia and CUNY, and there are $90 million in tax credits available to help smaller newsrooms continue hiring staff in the face of job losses due to AI. The relationship between writing and AI is something we will continue to navigate in the coming years and something I will continue to contemplate when ChatGPT-ing my discussion post due this Friday.

Artificial Intelligence via Bwarchive