bpcrd

We won multiple awards including RSA: https://www.rsaconference.com/library/press-release/reality-...

And we’ve published peer-reviewed research at top AI conferences E.g. CVPR, NeurIPS, ECCV, AAAI, Interspeech which are available at https://www.realitydefender.com/research

I completely agree! Please see the bottom of our post where we offer free access in two ways:

1) Email up to 50 files to yc@realitydefender.com, we’ll scan them for you, no setup required

2) 1-click add to Zoom/Teams (via Appstore) to try detection live in your own calls immediately

This is something for us to consider in the near future!

See my other response on this, but this is something we are actively protecting against.

This is something we're constantly updating, upgrading, iterating, and improving on. Every. Single. Day.

Whether it's introducing new models, deprecating old ones, or improving existing ones, there is an element of both staying current but also looking ahead at research. Many of the new models generating hyperreal content we catch on day one because they're based on existing technology and/or research.

It is in fact something we support with our enterprise clients and will roll out at a later date via Public API. Including the free tier.

Noted. Our marketing team only uses 640x480 CRTs and works exclusively in IE6, so will flag to them via Yahoo Messenger.

We'd rather not tip our hand on any/all techniques used to discern actual users from bad actors and those seeking to reverse engineer, but suffice to say we do have are methods (and plenty of them).

As noted elsewhere, we give confidence scores between 1-99%. We also use many different models for each modality for a more robust and complete answer with each scan, and each model has its own confidence score.

I understand this is in jest, but unfortunately AI generation tools more or less stopped the six-finger issue a couple of years ago. We are decidedly not a model used for the express detection of finger abnormalities, but a multi-model and multimodal detection platform — driven by our Public API (which you can try for free right now, btw) — which uses many different techniques to differentiate between content that is likely manipulated and likely not manipulated.

That said, neat gag.

Thank you. As an inference-based detection platform, our models go into every scan with the assumption that all files are both not the original/ground truth AND the files have been likely transcoded. We never say something is 0% or 100% fake because we don’t have that ground truth. That said, our award-winning models are able to say, with a confidence score of 1-99% — the higher being likely manipulated — which, in turn, is sent to the team using said detection to action as they will. Some use it as one of many signals to make an informed decision manually. Others have chosen to moderate or label accordingly. There are experts who’ve been called to testify on matters like this one, and some of them work on these very models.

As for synthetic content that is undetectable to the naked eye or ear, we are already there.

Thank you! We’ve been working on this since 2021 (and some of us a bit before that), and we’re reminded every day that we are ultimately working something that helps people on the macro and micro level. We want a world free of the malevolent uses of deepfakes for ourselves, our loved ones, and everyone beyond, and feel all should be privy to such protection.

I understand the appeal of hashing-based provenance techniques, though they’ve faced some significant challenges in practice that render them ineffective at best. While many model developers have explored these approaches with good intentions, we’ve seen that they can be easily circumvented or manipulated, particularly by sophisticated bad actors who may not follow voluntary standards.

We recognize that no detection solution is 100% accurate. There will be occasional false positives and negatives. That said, our independently verified an internal testing shows we’ve achieved the lowest error rates currently available for addressing deepfake detection.

I’d respectfully suggest that dismissing AI detection entirely might be premature, especially without hands-on evaluation. If you’re interested, I’d be happy to arrange a test environment where you could evaluate our solution’s performance firsthand and see how it might fit your specific use case.

We see who signs up for Reality Defender and instantly notice traffic patterns and other abnormalities that allow us to see if an account is in violation of terms of service. Also, our free tier is capped at 50 free scans a month which will not allow for said attackers to discern any tangible learnings or tactics they can use to bypass our detection models.

We've actually deployed to several Tier 1 banks and large enterprises already for various use-cases (verification, fraud detection, threat intelligence, etc.). The feedback that we've gotten so far is that our technology is high accuracy and a useful signal.

In terms of how our technology works, our research team has trained multiple detection models to look for specific visual and audio artifacts that the major generative models leave behind. These artifacts aren't perceptible to the human eye / ear, but they are actually very detectable to computer vision and audio models.

Each of these expert models gets combined into an ensemble system that weighs all the individual model outputs to reach a final conclusion.

We've got a rigorous process of collecting data from new generators, benchmarking them, and retraining our models when necessary. Often retrains aren't needed though, since our accuracy seems to transfer well across a given deepfake technique. So even if new diffusion or autoregressive models come out, for example, the artifacts tend to be similar and are still caught by our models.

I will say that our models are most heavily benchmarked on convincing audio/video/image impersonations of humans. While we can return results for items outside that scope, we've tended to focus training and benchmarking on human impersonations since that's typically the most dangerous risk for businesses.

So that's a caveat to keep in mind if you decide to try out our Developer Free Plan.

Great comments - Better results require more data, improving models, and onboarding new (different) models. All areas of focus for us!

There are advantages to both, as well as using existing blockchains (Ie. Bitcoin) to record small amounts of information via Merkle Trees.

Fun fact: The oldest blockchain pre-dates bitcoin....and started in the 1990's in the New York Times classified via a daily recorded hash value!

https://crypto.news/finding-the-oldest-blockchain-in-the-new...

We're up for the challenge! Send us your images / videos (ask@realitydefender.ai) and we'll send you the results!

We are optimizing compute requirements by isolating only the part of the media that requires scanning, so that we can pass along the cost savings to customers. At scale, the costs will low enough to scan all files just like anti-virus.

We agree. Dataset fidelity and bias are major concerns for publicly available datasets. For this reason we are working to develop programmatically created datasets along with anti-bias testing and policies.

Yes - Please contact us at ask@realitydefender.ai

Thank you! Please reach out to ask@realitydefender.ai :)

We’re fascinated by the potential applications of crypto in content provenance. In this example, a UGC video platform would need a way to initially determine the content hasn’t been manipulated before it’s signed, right? What about a live scenario where a deepfake mimicking an exec calls a manager to wire $10M (https://www.forbes.com/sites/thomasbrewster/2021/10/14/huge-...)

We totally recognize deepfake detection is a big & constantly evolving challenge, but we don't see that as a reason to cede the truth to bad actors :)

Great questions.

1 - We include multiple models for GAN and non-GAN related synthetic media.

2 - Models are only as good as the training data, and most training data breaks down in the real world because hackers have access to this same open source training data. So we create our own proprietary training data which we have automated, and we continuously update it based upon emerging deepfakes that we find in the wild.

3 - We target 95% accuracy with all public and proprietary training sets. And we continuously test and iterate both the data sets and the models.

4 - Our policies require a background check on all users to filter out bad actors. We additionally have technology safeguards in place to limit improper use.

We provide a probabilistic percentage result that is used by a trust and safety team to set limits (ie. flag or block content) so it is not a binary yes/no. We search for specific deepfake signatures and we explain what our results are identifying.

We are adding new roles to our careers page on www.realitydefender.ai but feel free to reach out to career@realitydefender.ai and we can discuss your interest!

We are working with few partners (including Microsoft) who are interested in integrating our solution. We are focused right on supporting large organizations (companies and governments) that need to scan user generated content at scale.

In 2017 deepfakes were pretty crude, today the avg person can’t tell a real face from a deepfake generated on a 5 year old iPhone. We expect the tech to continue moving in this direction. So, similar to anti-virus, we're approaching this problem with an iterative, multi-model solution that can evolve with the threat.

1- Each model looks for different deepfake signatures. By design, the models do not always agree, which is the goal. We are much more concerned with false negatives, and we target a min of 95% accuracy for our model of detection models. 2 - The challenge is educating users about results without requiring a PhD. Our platform is targeted for use by junior analysts in cyber security or trust and safety. 3 - This is a good suggestion. We are exploring how we can offer an unlimited plan that can cover our high compute costs (we run our multiple models in realtime).