Microsoft chief warns more deepfake threats could be coming soon

Pexels
(Image credit: Pexels)

When it comes to deepfakes, what we’ve seen so far is just the tip of the iceberg. In the near future, we won’t be certain if the person we’re speaking to on a video call is real, or an impostor, and crooks won’t have trouble creating an entire chronology of fake videos to support their claims, or trick people into believing the legitimacy of an offer or campaign.

These harrowing predictions come from Eric Horvitz, Microsoft’s chief science officer, in a new research paper, titled “On the horizon: Interactive and compositional deepfakes”.

Deepfakes are “photoshopped” videos, essentially. By using artificial intelligence (AI) and machine learning (ML), a threat actor is able to create a video of a person saying things that they never said. Now, according to Horvitz, crooks are ready to take it to the next level. Interactive deepfakes are just as you’d expect - real-time videos with which users can interact, which are, in reality, utterly fake.

Synthetic history

Compositional deepfakes, on the other hand, are described as “sets of deepfakes” designed to integrate over time with “observed, expected, and engineered world events to create persuasive synthetic histories.” 

“Synthetic histories can be constructed manually but may one day be guided by adversarial generative explanation (AGE) techniques,” Horvitz adds. 

He also says that in the near future, it will be almost impossible to distinguish fake videos and fake content from authentic ones: “In the absence of mitigations, interactive and compositional deepfakes threaten to move us closer to a post-epistemic world, where fact cannot be distinguished from fiction.”

This absence of mitigations stems from the fact that threat actors can pit artificial intelligence against analysis tools and develop deepfake content that is able to fool even the most advanced detection systems. 

“With this process at the foundation of deepfakes, neither pattern recognition techniques nor humans will be able to reliably recognize deepfakes,” Horvitz notes. 

So, next time a family member calls from abroad to ask for money to pay the rent, make sure it’s not a fraudster impersonating your loved ones. 

Via: VentureBeat

TOPICS

Sead is a seasoned freelance journalist based in Sarajevo, Bosnia and Herzegovina. He writes about IT (cloud, IoT, 5G, VPN) and cybersecurity (ransomware, data breaches, laws and regulations). In his career, spanning more than a decade, he’s written for numerous media outlets, including Al Jazeera Balkans. He’s also held several modules on content writing for Represent Communications.

Read more
Hands typing on a keyboard surrounded by security icons
Tackling the threat of deepfakes in the workplace
Honor's Deepfake Detection feature on an orange background
I hope this AI deepfake detection feature comes to more phones soon – but it needs one key upgrade to be truly useful
A deepfake image being created on a computer.
Deepfake scam calls are costing British victims hundreds each time - here's how to stay safe
AI deepfake faces
In a test, 2000 people were shown deepfake content, and only two of them managed to get a perfect score
A fish hook is lying across a computer keyboard, representing a phishing attack on a computer system
Everything you need to know about phishing
Concept art representing cybersecurity principles
Cybercriminals cashing in on holiday sales rush
Latest in Security
DeepSeek on a mobile phone
More US government departments ban controversial AI model DeepSeek
Ransomware
Fortinet firewall bugs are being targeted by LockBit ransomware hackers
Trojan
Microsoft warns of a devious new RAT malware which can avoid detection with apparent ease
NordProtect logo
Standalone identity theft protection from Nord Security is now available
A man holds a smartphone iPhone screen showing various social media apps including YouTube, TikTok, Facebook, Threads, Instagram and X
Ofcom cracks down on UK tech firms, will issue sanctions for illegal content
A fish hook is lying across a computer keyboard, representing a phishing attack on a computer system
These fake GitHub "security alerts" could actually let hackers hijack your account
Latest in News
Lego Pokemon
Pokemon and Lego announce the most electrifying collaboration of all time and I’m going to be first in line
Apple Watch app health
Apple Watch blood pressure monitoring tech revealed in patent
Using Zipped files and folders in Windows 11
Hidden clues suggest Microsoft is moving another part of Windows 11’s Control Panel to the Settings app – and this time it’s mouse options
an image of the Samsung Galaxy S24 Ultra
Finally! One UI 7 has a release date - here are the Samsung phones that’ll get it first
Google Cloud logo
Google to acquire cloud security platform Wiz in $32 billion deal
GIMP 3.0 interface from the website
Our favorite free photo editor finally got the update it deserves - and these are the top 5 features designers should know about