Yagizhan Atmaca retweeted
RT to help Simon raise awareness of prompt injection attacks in LLMs.
Feels a bit like the wild west of early computing, with computer viruses (now = malicious prompts hiding in web data/tools), and not well developed defenses (antivirus, or a lot more developed kernel/user space security paradigm where e.g. an agent is given very specific action types instead of the ability to run arbitrary bash scripts).
Conflicted because I want to be an early adopter of LLM agents in my personal computing but the wild west of possibility is holding me back.
If you use "AI agents" (LLMs that call tools) you need to be aware of the Lethal Trifecta
Any time you combine access to private data with exposure to untrusted content and the ability to externally communicate an attacker can trick the system into stealing your data!
ALT The lethal trifecta (diagram with three circles): access to private data, ability to externally communicate, exposure to untrusted content
Real-time deepfake pen test araclarimiz hazir ilgilenenler mail atabilir. [email protected]
Today, we are announcing PAGI Gen v1 Beta.
PAGI Gen is an img2img video framework. It makes possible for indie VFX artists to produce content of Hollywood quality. Try the beta now: pagi.ai/
Read more about: deepware.ai/unveiling-pagi-g…
Yagizhan Atmaca retweeted
Beyond fascinated by whatever voodoo @runwayml #Gen2 is conjuring. These are some outtakes from a short I'm releasing next week. Simulating water is some processor-intensive stuff. Thats why generating will replace rendering in future VFX #ai #aiart #generativevideo #generativeai
This video created with text-to-video technology. With just a few prompts. No need for expensive studios or actors - just a few prompts were enough to generate stunning visuals.
Pushed #gen2 again & made a movie trailer.
#aicinema is finally here!
Every shot made from text prompts, except one iconic shot you all know, done with #gen1
Made possible by @runwayml
@bazluhrmann your movies been a great inspiration! 😍🎞️🙏
Voices: @elevenlabs
#aianimation
Yagizhan Atmaca retweeted
Deepfakes have reached a new level.
AI will make you question the validity of everything.
Incredible example (with no post-processing) from @deepwareai
Yagizhan Atmaca retweeted
"The metaverse being unavailable to us at press time, we used Zoom" 😂
theverge.com/22588022/mark-z…
We are at the very beginning of biometric spoof attacks. Biometric identities for sale on the dark web. @mikko biometricupdate.com/202106/b…
After advanced masking tools, deepfakers are starting to use 3D models to improve missing angle issues! mrdeepfakes.com/forums/threa…
Kaspersky: “Protecting against deepfakes is going to require a drastic re-think of traditional approach.” Not just mp4 header classification :) which worked with %90+ acc in @kaggle dfdc challenge before the metadata leak fix. threatpost.com/deepfake-atta…
Yagizhan Atmaca retweeted
European MPs targeted by apparent deepfake video calls imitating Russian opposition. Always cautious calling deepfake w/out seeing actual video tho' real-time zoom facial puppeteering possible (thread). If so, escalation of political usages theguardian.com/world/2021/a… h/t @HenryAjder
The developer of DeepFaceLab working on something called #DeepFaceLive. This can be pretty convincing and dangerous compared to Avatarify and Metahuman. github.com/iperov/DeepFaceLi…
Fraudsters Use Deepfake Biometrics to Hack China’s Taxation System
findbiometrics.com/fraudster…
Yagizhan Atmaca retweeted
According to .@deepwareai , this video is 98% deepfake detected
scanner.deepware.ai/?fbclid=…
#whatshappeninginMyanmar
@freya_cole @shafiur @MayWongCNA @nslwin @MilkTeaMM_MTAM @cvdom2021 @ZtozM @somaticsoe @cape_diamond
The aim of #Myanmar #military showing a video of #Yangon Chief Minister's "confession" in trying to prove alleged corruption involving #AungSanSuuKyi has certainly backfired as it's trending on social media with many criticising how badly forged it was #WhatsHappeningInMyanmar
Yagizhan Atmaca retweeted
Today on Mar 23, at press conference, Military just released this video of U Phyo Min Thein with fake voice. The words and his lips weren't synchronized. So, the vid was analyzed with AI, it shows that the video is likely 75% deep fake.
#WhatsHappeningInMyanmar
Yagizhan Atmaca retweeted
China summons tech giants, including Alibaba and Tencent, over 'deep fakes' cna.asia/3cHI9t2
