Deepfake

The first well-known deepfake examples were presented to the public in the mid-2010s. For example, at first, deepfakes were used to create adult content using mainstream celebrities. Since then, advancements in AI algorithms have significantly improved deepfakes. 

For example, in 2016, deepfake technology was used to replicate Carrie Fisher and Peter Cushing onto other actors, with convincing results. Another popular example in recent times involves TikTok. There’s a popular account that posts deepfake videos of a person who looks like Tom Cruise.

You can detect deepfake by looking at simple red flags, such as unnatural eye blinks, mismatched earrings, or general asymmetries. 

Other suspicious signs to look out for include:

• Atypical body movement
• Strange textures and a glossy effect
• Unnatural colors, such as poor shadows or reflections
• Inconsistent lip-syncing paired with bad audio
• Suspicious videos or images that look poor-quality when zoomed in
• Out-of-place details like strands of hair, fuzzy jewelry, buttons, etc. 

It depends on the country you’re in and deepfake itself. In the US, there are no federal laws banning sexually explicit deepfakes, but several bills have been introduced in Congress, such as the DEFIANCE Act of 2024 and the AI Labeling Act of 2023. Despite that, neither has been approved by the committee. 

On the contrary, in the UK, under the Online Safety Act, which was passed in 2023, sharing deepfakes is illegal. The law similarly makes it illegal to create sexually explicit deepfakes, even without the intention to share them. This is based on the fact that most deepfakes can cause distress to the victim and cause humiliation. 

Critics argue that deepfake technology can be harmful due to their potential to create hateful content. For example, deepfakes have been used in celebrity pornographic content, fake news, financial fraud, and even child abuse material. In general, this is a threat to democratic systems and politics since deepfakes have been used to fake news. 

In today’s geopolitical landscape, where a lot of tension is happening, disinformation is a way of propaganda that interferes with informed decision-making. For this reason, governments are continuing to issue new laws and regulations that will restrict the use of AI and deepfake technology.

Yes, there are businesses that use deepfake technology for commercial purposes. A famous example of such a use case is the FakeApp application that was launched in 2017, allowing users to easily create and share videos with swapped faces. In 2019, FakeApp was replaced by open-source alternatives, including Faceswap.

For example, in 2019, the Japanese AI company DataGrid created full-body deepfakes that generate people with the intent to use their figures for fashion and apparel. As of today, larger businesses have been using deepfake technology to create deepfake avatars and voices. An example of such a service provider is Synthesia, a media generation company that develops software to create AI-generated video content.