It’s that time of year again: election season! You already know what to expect when you flip on the TV. Get ready for a barrage of commercials, each candidate saying enough to get you to like them but nothing specific enough to which they must stay beholden should they win.
What you might not expect is for sensationalist election “news” to barge in uninvited on your screens. Fake news – or exaggerated or completely falsified articles claiming to be unbiased and factual journalism, often spread via social media – can pop up anytime and anywhere. This election season’s fake news machine will be different than previous years because of the emergence of mainstream artificial intelligence tools.
AI’s Role in Fake News Generation
Here are a few ways desperate zealots may use various AI tools to stir unease and spread misinformation around the upcoming election.
We’ve had time to learn and operate by the adage of “Don’t believe everything you read on the internet.” But now, thanks to deepfake, that lesson must extend to “Don’t believe everything you SEE on the internet.” Deepfake is the digital manipulation of a video or photo. The result often depicts a scene that never happened. At a quick glance, deepfakes can look very real! Some still look real after studying them for a few minutes.
People may use deepfake to paint a candidate in a bad light or to spread sensationalized false news reports. For example, a deepfake could make it look like a candidate flashed a rude hand gesture or show a candidate partying with controversial public figures.
AI Voice Synthesizers
According to McAfee’s Beware the Artificial Imposter report, it only takes three seconds of authentic audio and minimal effort to create a mimicked voice with 85% accuracy. When someone puts their mind to it and takes the time to hone the voice clone, they can achieve a 95% voice match to the real deal.
Well-known politicians have thousands of seconds’ worth of audio clips available to anyone on the internet, giving voice cloners plenty of samples to choose from. Fake news spreaders could employ AI voice generators to add an authentic-sounding talk track to a deepfake video or to fabricate a snappy and sleazy “hot mike” clip to share far and wide online.
AI Text Generators
Programs like ChatGPT and Bard can make anyone sound intelligent and eloquent. In the hands of rabble-rousers, AI text generation tools can create articles that sound almost professional enough to be real. Plus, AI allows people to churn out content quickly, meaning that people could spread dozens of fake news reports daily. The number of fake articles is only limited by the slight imagination necessary to write a short prompt.
How to Spot AI-assisted Fake News
Before you get tricked by a fake news report, here are some ways to spot a malicious use of AI intended to mislead your political leanings:
Distorted images. Fabricated images and videos aren’t perfect. If you look closely, you can often spot the difference between real and fake. For example, AI-created art often adds extra fingers or creates faces that look blurry.
Robotic voices. When someone claims an audio clip is legitimate, listen closely to the voice as it could be AI-generated. AI voice synthesizers give themselves away not when you listen to the recording as a whole, but when you break it down syllable by syllable. A lot of editing is usually involved in fine tuning a voice clone. AI voices often make awkward pauses, clip words short, or put unnatural emphasis in the wrong places. Remember, most politicians are expert public speakers, so genuine speeches are likely to sound professional and rehearsed.
Strong emotions. No doubt about it, politics touch some sensitive nerves; however, if you see a post or “news report” that makes you incredibly angry or very sad, step away. Similar to phishing emails that urge readers to act without thinking, fake news reports stir up a frenzy – manipulating your emotions instead of using facts – to sway your way of thinking.
Share Responsibly and Question Everything
Is what you’re reading or seeing or hearing too bizarre to be true? That means it probably isn’t. If you’re interested in learning more about a political topic you came across on social media, do a quick search to corroborate a story. Have a list of respected news establishments bookmarked to make it quick and easy to ensure the authenticity of a report.
If you encounter fake news, the best way you can interact with it is to ignore it. Or, in cases where the content is offensive or incendiary, you should report it. Even if the fake news is laughably off-base, it’s still best not to share it with your network, because that’s exactly what the original poster wants: For as many people as possible to see their fabricated stories. All it takes is for someone within your network to look at it too quickly, believe it, and then perpetuate the lies.
It’s great if you’re passionate about politics and the various issues on the ballot. Passion is a powerful driver of change. But this election season, try to focus on what unites us, not what divides us.
The post This Election Season, Be on the Lookout for AI-generated Fake News appeared first on McAfee Blog.