DeepFaking – the use and misuse of deep learning AI.
Creative AI technology
One of the significant uses of deep learning AI is it’s application in image processing. Deep learning is capable of classifying images, identifying elements in the image, and effectively describing them in a coherent pattern. It cuts through the creative rigmarole which involves lengthy stints of creating and editing content to arrive at a desired content outcome.
Deep learning has the ability to draw out intricate information from images. This allows it to be fruitfully applied and used in different applications. Some significant developments in use are,
- In Google search and image search
- With it’s human-like cognitive ability it can identify products and categories in an image.
- In training corporate management as well as doctors
- In the insurance sector, like in case of automobiles
- For enhanced effects in video games
But, the amazing technology of AI can be effectively and negatively used to create video content for nefarious purposes. Using AI software, an adept person can transpose photo or video dynamics onto a real photo or video to misrepresent and convey an intended message with malicious intent. This act of using AI to create phony video content is referred to as deep faking.
Deepfaking is a convenient portmanteau that describes the significant technology of deep learning and it’s application to create altered or fake media content.
Till the emergence of the pathbreaking deep learning AI technology, an adept and qualified image editor required considerable experience and skills in order to achieve a desired altered image output.
This responsibility has now been assumed by AI algorithms. Therefore, agenda-driven individuals can easily manipulate images and videos using ready software.
Deepfaking is, for the most part, used for pranking. However the increasingly convincing outputs of the AI-enabled practice entering mainstream negative use is a visible threat along with the likes of 5G. It can precipitate serious compromises in people’s societal responses.
A good deepfake job can easily mislead people into accepting misrepresentations of people, situations, and entities. Further, it can create a lack of trust in people as well as in the genuineness of video content.
Sometimes, referred to as AI-generated mock media or synthetic content, it appears less menacing in 3D and other applications.
A decent learner by using the right software can create altered effects which can almost rival professionally rendered graphics imagery in a movie. Deepfaking outcomes are very realistic. Therefore, there already is serious concern about how doctored images laced with speech can harm individuals and important organisations.
A refined and realistic deepfake is the result of an ongoing, engaging interplay between the discriminating and enabling functions of AI. A constant interpretation of image elements is made available to the generator. It tries to create a realistic representation of the target output.
Till the desired outcome is transposed on the target image, the two algorithms continue interpretation and upgrading the alterations. Two genuine and interesting software sources for implementing this technology are FakeApp and DeepfaceApp.
Real-life instances of misuse
There have been a number of misuse cases. These have caused the affected person distress and have also managed to temporarily sway the perception of people and situations.
An impactful instance involved the use of After effects CC by Buzzfeed. The deepfake video involved the transposing of actor Peele’s face on Obama’s and incorporating the former’s mouth movements making it realistically appear as though Obama was speaking.
Another famous or infamous incident involved the projecting of Chinese actor Wang Yi onto a character of a popular television serial. The deepfake became immensely viral attracting 240 million views. Reaction to this one evoked an apology from the social perpetrator who stated it was a prank without intending harm to anyone. However there are others like the one involving Obama which are examples of how it can seriously affect important political figures at crucial times if AI is misused to create deepfake content. The ability of deepfake content to seriously impact a political or social context can create meaningful uncertainty.
On the other hand, this evolving AI technology provides massive substance and relevance to mainstream media like movies, animation, TV shows, learning programs, and corporate presentations, which is a major takeaway.