top of page
Search

The Threatening Duo of Deepfakes & Drones

  • Writer: Aastha Thakker
    Aastha Thakker
  • Oct 28, 2025
  • 4 min read

Hey everyone!


Last week, I presented on the topic “Threats!”. If you are from cyber security background, you will have an idea about the buzzwords like threats, risks, vulnerabilities, attacks etc.

Today, let’s take a different approach for this blog i.e. PPT method. Do share your reviews, suggestions, and questions at the end.


The flow of this blog will start with a basic introduction to threats, with a major focus on AI based threats and UAV threats.



What is a Threat?

A threat is something that poses danger. Let me give you a very simple example from real life. Imagine you are walking on a road- the act of walking is a vulnerability (weakness), the chance of you being hit by a vehicle is a threat (danger) and chance of your hand or leg being fractured is a risk — It's proportional to impact.


So, talking about threats, what are the types of threats?

These are the basic threats which I assume you might know & links are provided for further clarification. If you have any doubts, feel free to reach out on LinkedIn.


Now, let’s shift our focus to AI based threats (deepfake) and UAV threats.



In recent times, you may have heard about deepfake. It’s a technology that manipulates a media (images, audio, videos) to replace one person’s likeness or characteristics, convincingly, with that of another’s. Some recent examples include videos of the famous actor Rashmika and cricketer Sachin Tendulkar.


Here is another example; zoom in & mark the differences in the smile curve.
Here is another example; zoom in & mark the differences in the smile curve.

How does deepfake work?

It’s very complex, difficult and time-consuming process, but I will cover the surface.


Firstly, for creating any model we need a vast data pool. Through that big data we teach our models. Think of it like learning from many examples & data.


Two models are involved: Generator and Discriminator. Generator is used to create new images and Discriminator acts as a professor, detecting if the media is deepfaked. If the discriminator fails to detect deepfake then we are good to go otherwise not.


Then comes a crucial step, landmarking technique. In this technique images are landmarked with red dots and blue boxes. Minor error or negligence in this phase can impact the refinement process.

Audio synthesis becomes challenging, especially in videos where we have to manage lip sync along with text-to-speech audio conversion.

Finally, we have to do post processing and refinement to correct any errors before finalization.


Now, let's see significant threats of deepfake.


  1. Political Manipulation: Deepfake is used to create a deceptive media to influence political speeches or public opinion.

  2. Deepfake Pornography: This can even lead to psychological distress and society pressure.

  3. Financial Scams: This can lead to financial scams and frauds on individual or organization.

  4. Fake news & misinformation: Spreading misinformation can even cause havoc under different situations.

  5. Voice deepfakes fraud: This can lead to advanced social engineering attacks.



Thinking about the solutions, how can we detect or prevent deepfake from causing harm or spreading misinformation?

  1. Metadata Analysis: Checking timestamps and geolocation of the media.

  2. Consistent Lightning & background check: Verification and alertness regarding the lightning and shadows can prevent deepfake threats.

  3. Media authentication Platforms: Platforms like Truepic and many more should be used to check the authenticity of a media.

  4. Deepfake detection Tools: Tools like deepware scanner, microsoft’s video authenticator etc should be used to check the media.

  5. Biometric Authentication: Fingerprints or facial recognition like Apple’s face ID or Samsung’s fingerprint scanners can be used to verify unique biological traits.



Moving on to UAV (Unmanned aerial vehicle) Threats.

The drone you are saw above is UCAV (Unmanned combat aerial vehicle) which is used for military & national security purposes.


A) Consumer Drone: Used for personal use like DJI Mavic.

  1. Privacy Invasions: Unauthorized capturing private moments or spaces.

  2. Data Collection without consent: Collecting information without individual’s consent.

  3. Airspace violations: Unauthorized entry into restricted or private airspace.


B) Military Drone: Used for military & national security operations.

  1. Terrorism Threats: Bombarding and threating.

  2. GPS Spoofing: Manipulating GPS signals for deceptive purposes.


C) Agricultural Drone: used for crop monitoring & management.

  1. Crop surveillance: Unauthorized monitoring of agricultural activities.

  2. Crop damaging: Harming crops during surveillance.


D) Delivery Drone: Used for delivering packages like Amazon and Domino’s.

  1. Airspace violations

  2. Privacy concerns

  3. Package theft: Stealing packages.

  4. Collision Risks: With vehicle building or human.


E) Commercial Drone: Used for commercial activities and advert like Walmart.

  1. Criminal activity: Unlawful purpose.

  2. Espionage & corporate surveillance: Unauthorized monitoring of corporate activities.



Now checking the solution for drone threats, lets check the following measures.

  1. Drone Detection Systems : Technologies like radar or sensors to identify and track unauthorized drone activity.

  2. Jamming 7 signal disruptions : Disrupting communication signals, preventing them from functioning properly.

  3. Education & public awareness : Raising awareness about drones, its uses and threats.

  4. Cyber security measures : Implementing protective measures to secure channels & prevent unauthorized access or control of drones.

  5. Drones & Surveillance response : Improving protocols and strategies for incident response plan.


Comments

bottom of page