Kevin Systrom and Mike Krieger's brainchild, the Artifact news app, has raised the bar in the war against clickbait. Last month, they offered users the ability to flag misleading articles. Now, they're going one step further by using advanced AI technology to rewrite clickbait headlines.
Once an article gets flagged as clickbait, it triggers the GPT-4 model to transform the sensationalistic title into something more genuine. Originally, flagged articles underwent a manual review process by the Artifact team before earning the 'clickbait' tag.
In a major update, if the Artifact team validates an article as clickbait, AI will automatically revise the headline for all users, visible with a star icon. This new feature brings forth the promise of a 'clickbait-free' news environment.
Not stopping there, Artifact is also developing an automated clickbait detection system, eliminating the need for manual reporting. Once active, this feature will automatically detect and modify misleading headlines.
By launching these features, Artifact is voluntarily embracing a hefty editorial responsibility. This includes accurate identification of clickbait articles and employing AI to generate non-misleading headlines. However, the company isn't deterred by these challenges.
During a TechCrunch interview in March, Systrom acknowledged the editorial responsibility embedded in running a news app. According to him, every step from algorithm development, training data selection, data inclusion/exclusion, and objective weighing involves an editorial judgment.
