AI EDUCATION: What is a Deepfake?


Deepfakes.  We’ve all heard of them in the news and we all know they are being used more frequently and with some potentially troubling consequences.  Let’s delve into what these are, explore some recent examples and ponder how they might be prevented (or at least controlled) in the future.

Deepfakes are synthetic media, including images, videos, and audio, generated by artificial intelligence (AI) technology that portray something that does not exist in reality or events that have never occurred. The term “deepfake” combines “deep,” taken from AI deep-learning technology (a type of machine learning that involves multiple levels of processing), and “fake,” addressing that the content is not real.

Deepfakes can convincingly depict real (and often influential) individuals saying or doing things they never did, resulting in misleading content that can profoundly influence public opinion. This technology offers great benefits in the entertainment industry, but when abused for political manipulation, the capability of deepfakes to fabricate convincing disinformation could result in voter abstention, swaying elections, societal polarization, discrediting public figures, or even inciting geopolitical tensions.

Recent Examples of Deepfakes

Deepfakes have been used in various ways recently. For instance, a deepfake video “resurrected” the late President Suharto during the recent Indonesian election. This was ostensibly to encourage people to vote, but it was accused of being propaganda because it was produced by the political party that he led.

In another example, a robocall made to thousands of New Hampshire residents ahead of the state’s presidential primary in January featured an artificial-intelligence-generated dupe of President Joe Biden’s voice urging voters not to bother with the primary. NBC News later reported the audio was a complete fabrication created by a New Orleans street magician who said he was hired by a Democratic operative.

Many will have heard of the unfortunate deepfake of Taylor Swift in a non-consensual sexually explicit pose with the Kansas City Chiefs, and this type of deepfake has resulted in outrage (appropriately), lawsuits and even legislation (an AI replication prevention law recently passed in TN) to protect individuals from this type of exploitation.

Finally – on a lighter note – the AI image creation software Midjourney was used in 2023 to depict the pope in a designer puffer jacket (dispayed at the top of our post).  We personally thought this was rather funny and ironic – but – indeed, it may not have been deemed so amusing to many devout Catholics.

Technologies to Prevent or Identify Deepfakes

As deepfakes become more sophisticated, new technologies are being developed to identify and prevent them. Biometric technology provides a way to verify identities and it’s also becoming more sophisticated – moving beyond fingerprints and facial recognition to help people and organizations work out if the identity being verified even belongs to a real, live person.

Multiple technology-based detection systems already exist today. Using machine learning, neural networks, and forensic analysis, these systems can analyze digital content for inconsistencies typically associated with deepfakes. Forensic methods that examine facial manipulation can be used to verify the authenticity of a piece of content.

Stanford University has developed AI tools that can detect the lip-synching processes that are frequently used to put words never spoken into the mouths. Big tech organizations, including Microsoft, have developed toolkits to keep families safe online.

In conclusion, while deepfakes pose a significant challenge in the digital age, advancements in technology are providing us with the tools to combat this threat. As we continue to innovate and adapt, we can hope to stay one step ahead in the ongoing battle against deepfakes.

Publisher’s Note – for some additional insight into deepfakes, check out this week’s “AI & Finance | News for the Week” column from thought leader Chris Robbins.  

ChatGPT and DWN Staff

Article Sources

  1. Deepfake Definition & Meaning – Merriam-Webster
  2. Deepfake | History & Facts | Britannica
  3. 4 ways to future-proof against deepfakes in 2024 and beyond
  4. Deepfakes are still new, but 2024 could be the year they have an impact …
  5. States are cracking down on deepfakes ahead of the 2024 election
  6. New technologies are helping to identify sophisticated AI deepfakes …
  7. The US is drafting new laws to protect against AI-generated deepfakes
  8. deepfake | Meaning & Origin |
  9. Tackling deepfakes ‘has turned into an arms race’
  10. Deepfakes are still new, but 2024 could be the year they have an impact on elections
  11. Georgia lawmakers are using an AI deepfake video to try to ban political deepfakes
  12. Deepfakes, explained | MIT Sloan
  13. AI Tools Can Detect Deepfakes, But for How Long? | PCMag