Protect AI raised $13.5 million in a seed-funding round co-led by Acrew Capital and a group of other companies. Ian Swanson said that the capital will be put toward product development and customer outreach.
One of the few security companies focused solely on developing tools to defend artificial intelligence systems and machine learning models from exploits is Protect Artificial Intelligence. The product suite aims to help developers identify and fix security vulnerabilities in machine learning at different stages of the life cycle.
As machine learning models usage grows more and more in production use cases, we see artificial intelligence builders needing products and solutions to make their systems more secure. We have found unique exploits and provided tools to reduce risk.
About a year ago, Swanson co-launched Protect Artificial Intelligence with Daryan Dehghanpisheh and Badar Ahmad. At Amazon Web Services, Dehghanpisheh was the global leader for machine learning solution architects and Swanson was the worldwide leader in the Artificial Intelligence and Machine Learning side of the business. While working at DataScience.com, which was acquired byOracle, Ahmed became acquainted with Swanson. The VP of artificial intelligence and machine learning at Oracle was the one who worked with the two men.
Jupyter Notebook is a popular digital notebook tool used by data scientists in the artificial intelligence community. There were more than 2.5 million Jupyter notebooks in use at the time of the report, a number that has almost certainly climbed since then. Jupyter notebooks contain all the code, libraries and frameworks needed to train, run and test an artificial intelligence system.
What kind of problematic elements might be contained in the notebook? Swanson suggests internal use of credentials. Defense Newbery looks for personally identifiable information and open source code with a nonpermissive license that might prevent it from being used in a commercial system
Jupyter notebooks are usually used as scratch pads rather than production environments. Dark Reading found that less than 1% of instances of Jupyter Notebook on the public web are configured for open access. The exploits aren't just theoretical. A method that could allow an attacker to run any code on a victim's notebook across accounts on Amazon's fully managed machine learning service was discovered last December.
Aqua Security has found that Jupyter notebooks are vulnerable to hacking. The majority of businesses don't have the right tools in place to secure their machine learning models.
It's too early to sound the alarm. Despite a report predicting an increase in cyberattacks through the end of the year, there is no proof that attacks are happening at scale. The case is made that prevention is important.
Security code scanning solutions aren't compatible with Jupyter notebooks. These vulnerabilities, and many more, are due to a lack of focus and innovation from current cybersecurity solution providers.
Beyond Jupyter Notebooks, Protect Artificial Intelligence will work with a number of common artificial intelligence tools. It is free to begin, with paid options to come in the future.
Cybersecurity blind spots are created and prevented from being adequately understood and mitigated when machine learning is delivered at scales. Data sources, models, and software supply chain need to be hardened to meet increased governance, risk management, and compliance requirements. It helps enterprises of all sizes meet today's and tomorrow's unique, emerging and increasing requirements for a safer, more secure artificial intelligence powered digital experience.
That is promising a great deal. Protect is able to enter a market with few competitors. There is a company called Resistant Artificial Intelligence that is developing systems to protect against automated attacks.
ProtectAI isn't revealing how many customers it has. Finance, healthcare and life sciences are just some of the industries that the company has secured enterprise in the Fortune 500.
The funding will be used to add additional team members in software development, engineering, security and go-to-market roles over the course of the next five years. Several years of cash runway are available to advance this field.
All Rights Reserved © 2024
