Long championed as a transformational force that will disrupt everything, artificial intelligence has most consistently disrupted hypers — and not much else.

A few newsrooms use systems like Heliograf, The Washington Post’s tool for crafting machine-generated stories. Heliograf helps cover stories that otherwise had too small an audience to warrant coverage — like every individual House race in the 2018 midterm or sporting events. Machine-generated stories free up human reporters to do the stories only humans can, but they haven’t reached a wide enough scale to change the way most newsrooms’ approach coverage.

Even that level of machine-generated stories has been beyond the ability of most newsrooms and too complicated for most individual journalists. Journalists like John Keefe’s Quartz AI Studio and the work done at the Post by the Arc team seek to democratize the use of AI. More achievable projects — like feeding reporters story tips, automated tracking of data to uncover fraud, and scripts to keep published stories up-to-date — can make work easier for journalists.

Until now, AI was primarily limited to article recommendations and simple stories, but 2019 will finally see the widespread adoption of useful AI — and not just in large newsrooms like the Post. AI that helps our audience by adding contextual background information and aids journalists with story composition will be everywhere, allowing AI to finally fulfill its potential in media.

The promise of AI will be realized once non-technical reporters and editors benefit from AI in their existing jobs — whether they know AI helped or not. In 2019, we’ll see the beginning of that transformation.