Checkfor.ai | Application Preview
-
General Information
Business Registration Number: 93-3260011
Location: Brooklyn, NY, United States
Length of Operation: 1-5
Number of Employees: 1-10 Employees
Annual Gross Income: Less than $100k
Annual Gross Expense: Less than $100k
Open to Loans: NO
-
Funding Usage
We would like to use the funding to build out a world-class research team to further our development of the best-in-class AI text detection software for enterprises and government. We will primarily use the funding for (1) covering the operational costs of R&D, which is primarily cloud compute spending to train large neural networks to perform the detection, and (2) hiring a machine learning researcher to accelerate our pace of innovation.
-
Business Plan
Our long-term product vision is to develop a comprehensive authentication solution to identify and respond to AI-generated spam and fraud tailored to industry use cases and expand to other modalities, such as audio, images and videos. With additional capital, we would be able to massively scale our technology to keep up with increasingly advanced AI and solve the problem long-term. We need to stay ahead on the research front due to the increasing sophistication of adversaries. So far, we’ve started onboarding new customers who have already submitted over 100M documents to be checked by our solution. Through our traction, we believe we can scale and refine our trust & safety business to other online platforms while developing other security solutions. We recently released a technical report [https://checkfor.ai/blog/technical-report-february-2024] demonstrating that our technology is state-of-the-art compared to other AI detection commercial and open source methods, but this is just the tip of the iceberg. We do not view our company as simply AI detection; our business will grow into a next generation cybersecurity company as large language models become widely adopted across industries and across the globe. Just as people have learned to use computing technology for good, bad actors have used computers to manipulate and deceive at scale that had been previously unseen. LLMs are the next frontier: bad actors are already using LLM technology to deceive at even greater scale, causing unprecedented harm. Assuming that AGI is right around the corner, the internet needs an industry-standard protection layer for any form of content generation platform, and current tools are completely unequipped to handle this new category of deception at scale.
-
Self Identified Competition
GPTZero is a widely used AI detection tool primarily used by educators to check for plagiarism. Hive AI is a content moderation company that provides a suite of AI models and products to enterprises. Reality Defender is another startup that is attempting to solve enterprise deepfake detection, although their primary modalities are audio and video currently. Due to our team’s extensive experience training large deep learning models for industry (Tesla Autopilot, Google Ads), our model to detect AI-generated text is over 9 times better than the next closest competitor (GPTZero). More details on our model’s performance against the competition can be found in our team’s comprehensive technical report. [https://checkfor.ai/blog/technical-report-february-2024] Our team has both the research background and the practical industry experience to reliably train large-scale neural networks. Additionally, our focus on Internet content differentiates us from other AI detection startups that primarily focus on education. Given that there is a wealth of human produced user-generated content on the Internet, we believe that the problem reduces to a matter of who can most effectively develop algorithms that can search and learn patterns in vast amounts of data. We believe that our team’s experience working on similar problems makes us the best-equipped people to solve this problem long-term.
-
Contact Applicant
Subscribe to our Administrator Dashboard to gain full access to this application. Learn More
