“Since then, we have seen the fact-checking community in India grow to the point that India is actually the country that has the most fact-checkers in the world certified by the International Fact-Checking Network (IFCN),” she said. “We want to make sure the fact checks that they produce appear on our products. As a result, you can actually see fact checks in Google Search, Google News, and YouTube in both English and Hindi.”
Liu said that journalists are at the forefront of fighting misinformation and that with this in mind, the Google News Initiative (GNI) India Training Network was established. She said it’s focused on bringing the skills of fact-checking, digital safety and other issues to newsrooms where working journalists can teach others.
“We have 250 trainers across the country who can train in more than 10 languages. And since we started this program in 2018, we’ve trained 38,000 journalists, media educators, journalism students and fact-checkers across the country,” she said. “We are very pleased to expand this program this year to include new topics related to fact-checking and combating misinformation.”
She said topics like using data to improve fact-checking and how to check issues like climate change are among some topics to be covered in upcoming programs over the next few months.
On the policy front, Clement Wolf, Senior Public Policy Manager, Information Integrity at Google, said the company has policies in place that cover a range of harmful content and behavior. This includes, but is not limited to, hate speech about harassment and variations of misinformation and disinformation.
Discover the stories of your interest
“We evolve these policies over time to respond to new threats or new opportunities to do a good job,” he said. “Enforcing these policies at scale is one of the challenges of running platforms like ours.”
Google and the numerous platforms it operates, such as YouTube, are dependent on reports or notices that normal users or others identify as violating guidelines. Wolf said the company’s machine learning systems help him do this work at scale, but stressed that context is crucial when evaluating a piece of content for misinformation or disinformation.
“We really rely on this complementarity between human reviewers who have been trained and understood the nuances of the policy, and machine learning systems whose job it is to improve content for these people to review. And of course, the result of the checks informs the machine learning systems so they can do better over time.”
Wolf said Google removed over 3.4 billion ads in 2021 and took action on nearly 4 million channels in the last quarter of 2021. “We have policies for each of our services,” he explained. “These policies vary from service to service. Although we all look at the same harms across different services, those services don’t have the same purposes or user expectations, and so we may respond differently to those harms across those services.”
