“Deepfake:” it’s a stupid name, but a smart problem. In the age of cyber-security importance, an often-overlooked aspect of being online is the validity of web content. Deepfakes utilize machine-learning and artificial intelligence (AI) to, in certain scary situations, manipulate the pixels of the face to match a generated voice, enabling the person behind the information attack to make it seem as though any trusted figure is really telling you anything the attacker wants; It turns into a recipe so realistic that it’s scary.

Though it’s perhaps funny as an office gag to watch celebrities sing happy birthday to a colleague, or insert themselves into blockbuster films with stunning realism, deepfakes are a significant threat to your online presence. It may be easy to fact-check a Facebook friend who likes to post wild statements by simply searching for the correct answer, but when you see a trusted public figure making an outlandish remark, sometimes going against what you’ve known to be true for years, how can you be sure you’re getting the right information? How can you protect yourself against these hyper-realistic, machine learning algorithms manipulated by people who don’t have your best intentions in mind?

As with anything online, find the credible source of information. Seeing a video posted to YouTube or Facebook is not credible, first-hand documentation. First-hand information comes directly from the person or agency in charge of the findings. There’s no easy way of telling how many times something has been manipulated or shared, or even where it originated, on social media sites. Always look for secure sites that use the “https://” prefix when possible, look for the agency or group that originally published their findings, or look directly to political candidates’ campaign websites to find what they say their platforms really are, for example.

So how can one be 100% certain online? Unfortunately, one can’t. But remembering to stick to peer-reviewed studies, secure sites and firsthand information, as well as major trusted news sources is the safe way to go for now. As machine-learning becomes even more powerful, it may be possible to alter live-streams of public broadcasts, make fake news even more realistic, and potentially even alter history as seen on the web. In short, don’t dole out your trust easily, and if you find your system compromised due to a deepfake or any other scam, call Granite and we might be able to help. As always, stay safe out there, and thanks for reading.