AI and security a growing problem

Security will be one of the biggest challenges in deploying artificial intelligence. That was starting point for the first keynote at the machine learning conference ICML (International Conference on Machine Learning) in Stockholm.

The keynote was held by Dawn Song, CEO of Oasis Labs, and Professor of Computer Science at University of California, Berkeley, ranked as the most cited scholar in computer security. Her message to the 5,000 attendees at the conference was that when artificial intelligence becomes more capable the threats will be even stronger.

“This requires a community effort. Let’s tackle the big challenges together,” Dawn Song said.

A growing issue

AI and security is a growing issue. Just a couple of days before her lecture at ICML the Oasis Labs, a startup Dawn Song is CEO of, announced in a press release it has raised $45 million from investors. Founded by a team of academic and entrepreneurial leaders, Oasis Labs is designing a platform to overcome performance, security and privacy limitations that have hampered blockchain adoption to date. The company has garnered investment support from a wide spectrum of the blockchain ecosystem including venture capital, crypto funds and hardware manufacturers.

The Oasis platform embodies two core values. The first is bringing privacy and security protections previously not possible on a blockchain. The second is delivering processing capabilities that have orders of magnitude greater performance than existing systems, even for complex applications. This combination enables sophisticated, computationally-intensive and privacy-sensitive services such as machine learning on the blockchain.

More responsible ways to share data

“It’s exciting to see talented people like Dawn and her team working on ways to transition the internet away from data silos and towards a world with more responsible ways to share and own your data,” said Fred Ehrsam, co-founder of Coinbase and Oasis Labs investor in the press release.

“Blockchains are poised to revolutionize much of the way we live, but many developers and organizations have understandable concerns about performance and privacy limitations that are currently hindering their ability to embrace the technology,” said Dr. Dawn Song.

“The Oasis platform aims to help users leverage and reclaim control over their data, and at the same time deliver superior performance and privacy capabilities. Our goal is to build the scalable and secure decentralized internet that puts users first.”

Challenges at the intersection of AI and security

In her keynote at ICML, Dawn Song talked about challenges and exciting new opportunities at the intersection of AI and security, how AI and deep learning can enable better security, and how security can enable better AI.

She also gave an overview on challenges and new techniques to enable privacy-preserving machine learning and recent project on confidentiality-preserving smart contracts towards democratization of AI.

About Oasis Labs

Oasis Labs is building a privacy-first, high-performance cloud computing platform on blockchain. Oasis’s integrated hardware-software technologies aim to enable cloud-scale, real-world applications, including machine learning, for the first time on blockchain. Oasis Labs is led by a team of academic and entrepreneurial leaders and backed by a diverse range of investors including a16zcrypto, Accel and Binance Labs.