If artificial intelligence is a solution for journalism, bias is its Achilles Heel

Geraldine Moriba
Dec 6, 2018 · 3 min read

Can artificial intelligence-assisted programs spot the bias in reporting and flag it before it is published? That’s what I want to find out.

What’s wrong with the description in the Tweet below?

Would it have been written differently if it were about a border with another nation? Something like this… Tear gas was used against unarmed asylum seekers, including mothers and children.

As a part of my JSK Journalism fellowship at Stanford, I am researching ways to use artificial intelligence (AI) to identify biases in journalists. We’ve been slow to recognize and adopt the potential benefits of AI. Rather than building our own breaking news tools, we let Facebook and Twitter dominate our domain. Even though we are at the center of information gathering, we left it to Google to lead the way. In desperation, we even turned to Snapchat to reach a younger demographic —and failed. These non-news corporations have no obligation to follow journalism standards. They are designed to surface information using algorithms that measure which stories attract the most attention. Consequently, the most attention-grabbing content, regardless of its news value, accuracy or veracity, rises to the top. It’s quantity over quality. This relinquishing of journalistic responsibility is happening again right now in the race to find solutions that identify fake videos, verify facts, analyze data and single out automated bots.

“We have such a unique industry and we are so used to change every day, and yet we still cannot seem to innovate our way out of anything.” Raju Narisetti, Columbia Journalism Professor of Professional Practice and Director of the Knight-Bagehot Fellowship in Economics and Business Journalism

We are also missing a valuable opportunity to find ways to use AI to monitor our own biases as journalists. Selecting, editing and positioning news stories are entirely subjective processes. News organizations are rooted in cultural competition, crisis amplification and the shortsightedness that results from focusing on so-called “core” audiences. Editorial tools that increase journalistic self-awareness and accountability are needed.

Five reasons journalists should use AI to identify their own biases…

  • To better report on the broad spectrum of viewpoints, lived experiences and beliefs
  • To improve the quality of reporting and increase trust by being more transparent about the patterns of story coverage and perspectives
  • To resist news manipulation and misinformation
  • To rebuild trust in reporting and increase credibility
  • To provide the affirmation of dignity for everyone

Information gathering and distribution cannot be left to the algorithmic black boxes controlled by the Internet giants, profiteers or political opportunists. Journalists should move in front of finding ways that AI can benefit information gathering, especially now as trust in journalism is being eroded and alternative information pipelines are proliferating.

Ways that AI can be used to check biases

  • Developing tools to check the biases of journalists before publishing. This includes all editorial content, images, and video. As well as, monitoring who is doing the reporting and whose stories are being told.
  • Monitoring the biases and representation of the programmers who build AI algorithms. There are intentional and unintentional biases embedded in the datasets with which AI is trained.
  • Establishing AI standards around safety, privacy and ethics.
  • Supporting access to accurate and bias-free data sets from all levels of government, academic institutions and private corporations.
  • Investing in form-agnostic journalism that lets news consumers interact with information in more personalized ways to reach their own conclusions and foster critical thinking.

“Journalists shouldn’t aim to be neutral, we should aim to be truthful. That’s what we do as journalists. Look for the truth.” Christiane Amanpour, CNN Chief International Anchor and Host

This mission hasn’t changed. It shouldn’t.

Ultimately, best use practices of AI for journalism are twofold. These algorithms can be used to make newsrooms more competitive in the information space. They can also be used to challenge bias in journalists and foster self-reflective journalism.

If you have ideas, questions, or feedback on identifying AI solutions to monitor and challenge biases — let me know. I am open to partners and collaborators. I admit, from the jump, that I have a ton to learn, but I believe that this exploration is crucial.

JSK Class of 2019

Insights and updates from members of the John S. Knight Journalism Fellowships Class of 2019 at Stanford University

Geraldine Moriba

Written by

JSK Journalism Fellow at Stanford, Class of 2019. Emmy Winning Documentary Filmmaker. Writer.

JSK Class of 2019

Insights and updates from members of the John S. Knight Journalism Fellowships Class of 2019 at Stanford University

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade