Event Coverage Highlight
Newsroom AI Experts Present Tools and Best Practices for International Journalists
by Chad Bouchard
New artificial intelligence tools are poised to transform newsrooms and amplify their reporting power, and are essential weapons against bad actors waging disinformation campaigns that exploit A.I. tools of their own.
On Oct. 16, the OPC hosted a virtual panel discussion with experts using A.I. within their newsrooms.
Phoebe Connelly, senior editor, A.I. Strategy and Innovation at The Washington Post, moderated. She asked panelists about their approaches for navigating ethical issues and how they direct journalists to use their tools.
Rubina Madan Fillion, associate editorial director of A.I. Initiatives at The New York Times, said in a recent project, before journalists were given access to generative A.I. tools, they received training on standards and guard rails for how to use the technology.
“The most important one is that everything is done with human guidance and review. We don’t use A.I. to write articles,” she said. “Journalists are ultimately responsible for everything that we publish, and we want to make that really clear, to our audience but also to our journalists who are getting access to these tools. And we are also very interested in transparency.”
Fillion said a key feature of A.I. in newsrooms is its ability to process and organize large amounts of information. For example, you could use a smart spreadsheet tool to help investigative journalists figure out what to focus on in large data sets, and winnow out story leads. Reporters could take a list of 10,000 major political donors and distill it down to the 100 or so that are worth looking into more closely. While it would be time consuming for journalists to research thousands of people or organizations, A.I. can speed up that process and make it more manageable.
Aimee Rinehart, senior product manager A.I. strategy for The Associated Press, said the team at AP is continually evolving standards, not just for the newsroom but for the product team, evaluating the language models that they use and deciding when to shutter tools.
“It’s kind of an exciting time if you’re on a product team, because we’re definitely working very closely with everyone in the news operation.”
Rinehart talked about a recent pilot collaboration between the AP’s Local Lede program and AppliedXL that identifies anomalies and other helpful information from data in the Federal Registry. For example, in states with big questions around land use, like Alaska, Montana, Idaho or Hawaii, it can flag reporting tips that might otherwise be overlooked.
“Because if you’ve ever looked at the Federal Registry, it really does seem impenetrable,” she said. “This brings it into a news format so that people can better process what is at stake.”
Elyse Samuels, senior producer on The Washington Post’s Visual Forensics team, talked about how their team rigorously investigates the authenticity of video, audio and photographs, first with traditional means.
“While deep fakes and manipulated videos are getting much more advanced than even a few years ago, there are still visual anomalies that you can look for that make it quite obvious that it is in fact fake,” she said. “So we really do employ our journalists, or anyone on the internet engaging with this material, to act as you would in any other kind of situation, and bring that skepticism of asking – even the basics of the five w’s; who, what, where, when, why.”
She said they use geolocation techniques to confirm images were taken where they claim to be taken, comparing metadata and reliable sources such as Google Maps or other photos of the same location.
“It is about having a discerning eye and doing as much as you can to manually verify some of this material. But it definitely is a scary space. And I would love if there was a magic A.I. tool that could help you detect it.”
She added that there are some useful A.I. tools available. Fillion mentioned TrueMedia.org, a nonprofit organization offering tools that were recently used to combat political deepfakes in elections in India.
During the program, the panelists mentioned many other A.I. tools, policy statements and examples of projects, as well as podcast and newsletter recommendations. Explore the tables below for links to those resources.
AI Tools
Tool Name/Hyperlink | Description |
---|---|
ChatGPT | A chatbot that uses machine learning to generate human-like text and code in response to user prompts. |
NotebookLM | Note taking and research assistant powered by AI. |
TrueMedia | A tool for identifying political deepfakes in social media using AI. |
GPTZero | A free plagiarism checker that also identifies likely AI-generated material. |
Journalist’s Toolbox | AI tools and other resources for journalists. |
Google Pinpoint | A tool for analyzing large collections of documents. |
AI Newsroom Policy
Organization | Policy Page/Hyperlink |
---|---|
The Associated Press | AI guidance, terms added to AP Stylebook |
The Associated Press | Updates to generative AI standards |
The New York Times | Principles for Using Generative A․I․ in The Times’s Newsroom |
The Washington Post | AI Policy |
Examples of AI Use in Newsrooms
Organization | Article Title/Hyperlink |
---|---|
Medium | Teaching a Custom GPT to Read Audit Reports and Support Watchdog Journalism |
Semafor | Life during the Russian offensive, animated with AI |
The New York Times | Trump’s Speeches, Increasingly Angry and Rambling, Reignite the Question of Age |
The Washington Post | Republicans flood TV with misleading ads about immigration, border |
The Washington Post | How China extended its repression into an American city |
The Associated Press | AI experimentation is high risk, high reward for low-profile political campaigns |
The New York Times | Medium | How The New York Times Uses Machine Learning To Make Its Paywall Smarter |
The New York Times | Medium | How The New York Times Incorporates Editorial Judgement in Algorithms to Curate Home Screen Content |
The New York Times | How Stump Speeches by Harris and Trump Differ (and Don’t) |
Recommendations | Learn More
Resource Type | Name/Hyperlink |
---|---|
Podcast | Newsroom Robots |
Podcast | Hard Fork |
Newsletter | Axios AI+ |
Newsletter | Semafor Tech |
Newsletter | The Neuron |
Newsletter | Wonder Tools |
Newsletter | One Useful Thing |
Newsletter | Nieman Lab |
Organization | Online News Association (ONA) A.I. in Journalism Initiative |