Takes one to know one: New tool uses AI to spot text written by AI

Takes one to know one

A new tool uses AI to spot text written by AI

Jul 26, 2019

AI algorithms can generate text convincing enough to fool the average human—potentially providing a way to mass-produce fake news, bogus reviews, and phony social accounts. Thankfully, AI can now be used to identify fake text, too.
The news: Researchers from Harvard University and the MIT-IBM Watson AI Labhave developed a new tool for spotting text that has been generated using AI. Called the Giant Language Model Test Room (GLTR), it exploits the fact that AI text generators rely on statistical patterns in text, as opposed to the actual meaning of words and sentences. In other words, the tool can tell if the words you’re reading seem too predictable to have been written by a human hand.
The context: Misinformation is increasingly being automated, and the technology required to generate fake text and imagery is advancing fast. AI-powered tools such as this may become valuable weapons in the fight to catch fake news, deepfakes, and twitter bots.
Faking it: Researchers at OpenAI recently demonstrated an algorithm capable of dreaming up surprisingly realistic passages. They fed huge amounts of text into a large machine-learning model, which learned to pick up statistical patterns in those words. The Harvard team developed their tool using a version of the OpenAI code that was released publicly. 
How predictable: GLTR highlights words that are statistically likely to appear after the preceding word in the text. As shown in the passage above (from Infinite Jest), the most predictable words are green; less predictable are yellow and red; and least predictable are purple. When tested on snippets of text written by OpenAI’s algorithm, it finds a lot of predictability. Genuine news articles and scientific abstracts contain more surprises.
Mind and machine: The researchers behind GLTR carried out another experiment as well. They asked Harvard students to identify AI-generated text—first without the tool, and then with the help of its highlighting. The students were able to spot only half of all fakes on their own, but 72% when given the tool. “Our goal is to create human and AI collaboration systems,” says Sebastian Gehrmann, a PhD student involved in the work.
If you're interested, you can try it out for yourself.
https://www.technologyreview.com/f/614021/a-new-tool-uses-ai-to-spot-text-written-by-ai/


Comments

Post a Comment

Popular posts from this blog

BMW traps alleged thief by remotely locking him in car

BMW traps alleged thief by remotely locking him in car Stealer's Wheel? Seattle police department quotes "Watchmen" movie in a recap of the recent arrest. Tech Culture by Gael Fashingbauer Cooper December 4, 2016 5:00 PM PST It's maybe the most satisfying arrest we can imagine. Seattle police caught an alleged car thief by enlisting the help of car maker BMW to both track and then remotely lock the luckless criminal in the very car he was trying to steal. Jonah Spangenthal-Lee, deputy director of communications for the Seattle Police Department, posted a witty summary of the event on the SPD's blog on Wednesday. Turns out if you're inside a stolen car, it's perhaps not the best time to take a nap. "A car thief awoke from a sound slumber Sunday morning (Nov. 27) to find he had been remotely locked inside a stolen BMW, just as Seattle police officers were bearing down on him," Spangenthal-Lee wrote. The suspect found a ke...
Read more

Report: World’s 1st remote brain surgery via 5G network performed in China

World’s 1st remote brain surgery via 5G network performed in China Published time: 17 Mar, 2019 13:12 ·           A Chinese surgeon has performed the world’s first remote brain surgery using 5G technology, with the patient 3,000km away from the operating doctor. Dr. Ling Zhipei remotely implanted a neurostimulator into his patient’s brain on Saturday, Chinese state-run media  reports . The surgeon manipulated the instruments in the Beijing-based PLAGH hospital from a clinic subsidiary on the southern Hainan island, located 3,000km away. The surgery is said to have lasted three hours and ended successfully. The patient, suffering from Parkinson’s disease, is said to be feeling well after the pioneering operation. The doctor used a computer connected to the next-generation 5G network developed by Chinese tech giant Huawei. The new device enabled a near real-time connection, according to Dr. Ling.  “You barely feel that th...
Read more

New ATM's: withdraw money with veins in your finger

New cash machines: withdraw money with veins in your finger Cash machine technology that reads the pattern of finger veins is already available in Japan and Poland   By Telegraph Reporters 6:59PM BST 15 May 2014 Cash machines could soon be installed with devices that identify customers by reading the veins in their fingers. The technology is already being rolled out in Poland, where 1,730 cash machines will this year be installed with readers, negating the need for a debit card and Pin. Developed by Hitachi, the Japanese electronics firm, the machines read the patterns of the veins just below the surface of the skin on your finger using infra-red sensors. The light is partially absorbed by haemoglobin in the veins to capture a unique finger vein pattern profile, which is matched to a profile. The technology is used by Japanese banks and also in Turkey, offering “groundbreaking levels of accuracy and speed of authentication”, Hitachi said, which in t...
Read more