A startup has developed artificial intelligence software capable of detecting fake news, malware and problematic content.
For years the famous ‘fake news’ they circulate through the network in search of misinforming users to alter their opinion on political, economic, social or scientific issues, among others. It’s not rumors, it’s just misleading news that is sent quickly to reach more people, is anonymous, has no reliable source, does not expire and encourages the reader to spread it.
How many times have you received this type of hoax on WhatsApp or Facebook? In this regard, social networks have great responsibility because they are media that spread fake news more quickly. Therefore, Facebook has already included, in collaboration with the organization First Draft, a series of tips to detect ‘ fake news’. You can access through the ‘news’ tab of your wall or in suggestions.
To try to fix this problem, the startup AdVerif.ai it has developed a new algorithm that works with artificial intelligence (AI) software designed to detect fake news, malware and other types of misleading or problematic content. The startup, which launched the beta in November, is working with content platforms and ad networks in the US and Europe.
News verification with database
How does it work? The software analyzes the contents to detect suspicions that something is not right, such as too much capital letters in a headline or that they do not match what the text says. In addition, it is responsible for verifying the news with its database thousands of legitimate and false information that are updated weekly.
In this way, customers receive a report for each content that the software analyzes, where it indicates the probability that there is ‘fake news’, malicious content or any other problem. After testing AdVerif.ai, the software recognized the publication of ‘The Onion’ (a medium similar to The World Today) as a satire; identified when a Twitter account logo was not associated with the brand it corresponded to; and discovered that a ‘Natural News’ story with the headline ‘Evidence points to bitcoin as a psychological war designed by the US National Security Agency to launch a global digital currency’ was a ‘fake new’.
However, even if effective programmes are developed to detect false content, much remains to be done to combat them. In fact, in December 2016 experts in artificial intelligence developed the ‘Fake News Challenge ‘ platform with the aim of creating new technologies to combat the dissemination of false and misleading content.
It is very easy to make money by generating fake news thanks to the lack of supervision of the automation of large volumes of information. In this regard, the consulting firm Gartner maintains that by 2022 users will consume more false than true content, and there will be no technological capacity to combat it. However, there are more and more initiatives, such as AdVerif.ai, who fight this problem.
In the meantime, we recommend a series of guidelines to detect fake news:
- If the headlines are very eye-catching, appear in capital letters or with exclamation marks, doubt them.
- The URL it usually takes links from original media by imitating them.
- If the format it’s unusual and there are misspellings, it’s fake!
- Investigate the source. Most fake news is spread by unknown media or organizations.
- Check date because fake news usually has no expiration date.
- Verify information. See in reliable media whether this news has been published or not.
- Don’t forget that there are pages that are dedicated to spreading humorous content, such as The Today World, which has nothing to do with fake news.