Fact-based journalism vs fake news
Artificial intelligence as a safeguard against lies in real time.
In an age dominated by so-called fake news, artificial intelligence can help journalists verify the veracity of certain statements in real time, possibly during a live radio or TV broadcast.
On this front, an independent British organization is currently working on two products whose release is scheduled for 2018. In collaboration with Google, Full Fact (together with the Open Society and Omidyar) is developing "Live" and "Trends".
“Live”, made possible by Natural Language Processing, could be used during a live broadcast rally in which, for example, Donald Trump boasts (as he has actually done) that thanks to him “Fiat Chrysler Automobiles will leave Mexico to come back to the USA.” The product would highlight Trump’s statement with an on-screen text and indicate that it is misleading because - and this could be an explanation - while FCA has indeed stated it will transfer production of its pickup trucks from Mexico to the US, the company will continue to use its Saltillo plant to produce commercial vehicles instead of the Ram Heavy Duty trucks.
The second product, “Trends” instead monitors recurring statements and determines who is spreading them and why. This could allow those journalists who are responsible for verifying the statements’ accuracy to develop an idea on the motivations behind them and formulate specific requests.
Companies like the New York Times, the Guardian and the Economist are testing solutions to help them improve their article comment sections. Launched by Jigsaw, a Google offshoot, Perspective is a designed to defeat online harassment through machine learning. The product automatically identifies online insults and abuse and appears to be better than any blacklisting of words or faster than any human moderator. Although it is intended to create a safe place for online discussions, Perspective is not without critics. The automatic system could end up censoring or erasing certain expressions without understanding the context or taking away the freedom to say that "Donald Trump is a fool". To the critics, Jigsaw responded by saying that the tool is not meant to have the last word on which comments to publish and which not. But it can certainly help speed up the moderators’ work.