It’s taken a few years to get to where we are now in the AI industry, but the acceleration of its growth is more exponential than linear. With more data comes more data (comes more data), and so the rate of progression is increasing fast. So fast that many government ‘decision-makers’ agree that AI is not on a steady enough track to be moving at the rate it is. Around 51% of those business leaders claim that AI is gaining momentum fast in their industry, while a whole 37% also concur that it is moving ‘too [quick] for comfort.’ The top two potential risks the decision-makers decided on were cybersecurity breaches (63%) and privacy violations (42%) - which are not targeted invasions, but rather ones that could affect an entire population of users. This led to more than two-thirds of business officials admitting that it’s ‘difficult to stay on top of constantly evolving AI offerings’ (Dille).
Similar to fake news, artificial intelligence is also capable of creating fake audio, photos, and video by using ‘a form of artificial intelligence called deep learning [to create fake scenarios] ...hence the name deepfake.’ Creators of deepfakes range from research scientists to visual effects directors to porn producers. Deepfakes have been used in the past to show celebrities and politicians saying things that had never been said by those people, such as Mark Zuckerberg claiming that he had “total control of billions of people’s stolen data.” Developers can even use the technology so that AI becomes programmed to completely ‘make up’ fake faces- something the human brain isn’t even capable of doing. Fortunately- and ironically- AI also has existing detection systems for deepfakes. Another concern regarding deepfakes are shallowfakes, which are deepfakes but made without the advanced technology (or just presented in the wrong context). Not all deepfake technology are made with malicious intent, however- for example, some can be used as a ‘voice-clone’ for those who lose theirs to disease or to ‘resurrect a dead actor’ (or rather, just the actor’s image) so they could star in a new movie (Sample).
When looking at AI’s technological capabilities, the question comes up: how will AI change the fake news industry? While there are many AI systems out there being used to detect fake news, there are also others that are being programmed to create fake news. When AI is trusted with evaluating sources instead of people, racial and gender-related biases may be exposed (or, depending on the programmer, encouraged). This can result in restriction of usage on certain media platforms that can impact even political expression. Not only that, but if AI is designed to comb through news sources, vocabulary is key. And vocabulary changes quickly- what might be spotted one day as a medical term might be overused as a derogatory term the next. Is AI going to update itself every few years to learn to predict what vocabulary will shift into something offensive that could contribute to biased or fake news? It seems that people will have to rely on their own intelligence to analyze a news source- although at this point, there’s enough fake information out there that we might be seeing just as much ‘lookalike’ content as we do trustworthy (Susarla).
