Donald Trump single-handedly popularized the term “fake news”.Fake news is content, containing deliberate misinformation with the intent of misleading and misinforming readers. In recent years, most textual fake news stories were distributed via social media platforms like Facebook or custom URLs that looked like trustworthy news-outlets. Fake news can have a significant negative impact on society. For example, important elections, like Brexit, are decided on a small percent. Fake news can influence that small percentage, thus changing the outcome of the vote. At the time of writing, platforms like Facebook are taking measures to combat the distribution of fake news. On their platform, they have systems in place that can detect, flag or delete fake news stories (Facebook, 2017). Even though these measures are desirable first steps, they probably will not be enough. The fake news problem will only grow, and AI is to blame.
Breakthroughs like Deep Learning have put AI in the mainstream. The latter technology makes it possible for computers to solve problems on its own, rather than manually coding the solution to a problem. This enables products that previously were impossible to create, e.g. Adobe VoCo; a Photoshop for speech. This software can replicate human voices. Users type their input, choose a person they want to imitate, and the system will read back your input in that person’s voice. This can also be done with video through the software Face2Face. The software first learns, through a recording, how a person’s face moves and operates. The user can then make the person’s face to do virtually anything they want. Combine the previously named software together, and it is not hard to imagine how they can enlargen the problem of fake news. Just recently, thieves used a deepfake audio recording to trick a company’s chief executive into wiring $222,000. The executive in question, unfortunately, believed he the voice of the CEO (The Verge, 2019).
Even though these technologies can be used for, e.g. Hollywood, most of these technological innovations are bound to have undesired side effects. Though AI can and probably will increase the creation of fake news. It does not mean we should ban or restrict it. We cannot prevent the use of deepfakes for evil. We can, however, educate the general public of the existence of these tools and how they can corrupt our media.
What is your opinion on deepfakes? Are you scared of their potential to damage our media? Did you even know this was possible? Let me know in the comments below!
