Early Stage: Artificial intelligence vs. fake news
What it is: NewsBotAI, a Facebook Messenger bot built by two UC Berkeley students
Source: Marisa Kendall
What it does: Detects fake news
Why it’s cool: Made-up and misleading “news” reports with political agendas flooded social media during the recent presidential election, and continue to cause controversy months later — some critics have gone so far as to suggest fake news helped get President Donald Trump elected. Facebook recently rolled out new fact-checking mechanisms and ways to flag fake news on the site, but at least one study found those safeguards so far have been ineffective.
That’s where NewsBot comes in. The chat bot uses artificial intelligence to spot fake news, but it also can decipher political biases in legitimate news stories and outlets. To use it, you just send the chat bot a link to a news article and wait for its reply.
NewsBot’s creator, Ash Bhat, says he also wants his bot to help restore people’s trust in the reputable news outlets that have been tarnished by the fake news controversy.
It seemed to work as intended when I tested it out. When I sent NewsBot a link to a Dc Gazette article titled “Another person investigating the Clintons turns up dead,” it warned me that site tends to publish fake news (a verdict it accompanied with a sad-face emoji). When I fed it a Breitbart story titled “Hillary Clinton adds DNC to list of groups she blames for election defeat,” it ranked the story as legitimate, but right-leaning. The bot ranked the validity of mercurynews.com as “generally high.” And when I fed it three stories I’d written about Trump’s travel ban, it ranked them as “center,” “moderate with a slight left bias” and “left-leaning” — possibly because the last article included several strongly worded quotes from tech leaders opposed to the travel ban.
Where it stands: NewsBot launched in April and already has several thousand daily users. Bhat says the bot is 85 percent accurate, and is getting better every day.
Only in Silicon Valley:
With so many high-tech fitness trackers on the market, these days it seems like we barely have use for the traditional bathroom scale.
San Francisco-based ShapeScale hopes to change that. The company makes a smart scale that uses 3-D scanning technology to calculate not just your weight, but your body fat percentage, body composition and the measurements of each body part. Step onto the scale and a robotic arm revolves around you, taking hundreds of pictures that are then stitched together to create a 3-D image of your body. The scale connects to an app where you can see the image, which also indicates where you’ve been gaining or losing fat, and where you’ve been adding muscle.
You can preorder the device at shapescale.com for $299.99, plus a $9.99 monthly subscription.
Run the numbers:
Ever sent a text message you wished you could take back? You’re not the only one. More than 70 percent of people have done the same thing, according to a recent study of 1,000 American mobile users sponsored by On Second Thought, an app that lets you edit or delete mistaken text messages after sending them. According to the study, 39 percent of respondents upset someone close to them with a bad text, and 16 percent were bullied or faced professional consequences as a result. And 58 percent knew immediately after pressing send that they had made a mistake.
Hillary Clinton spoke last week at tech blog Recode’s Code Conference in Rancho Palos Verdes. When asked whether Twitter has been good or bad for society, she said:
“I think it has become victimized by deliberate efforts to shape the conversation, and push it towards conspiracies, lies, false information. And I think it’s the same problem that Facebook faces, that when you try to be all things to all people and you try to open up your platform so that people can come in, and you want to be influential because you expect people will actually tune you in and read and watch what you have, what do you do to try and contain the weaponization and manipulation of that information? I don’t think we know yet.”