Security & privacy

Using software to manipulate a video of President Obama speaking

AI researchers race to detect ‘deepfake’ videos: ‘We are outgunned’

06/13/19 Washington Post — EECS graduate student Shruti Agarwal and incoming professor Hany Farid argue that powerful new AI software has effectively democratized the creation of convincing “deepfake” videos, making it easier than ever to fabricate someone appearing to say or do something they didn't really do.

Berkeley links up with Facebook, but wants to see tech giant’s accountability

05/17/19 — Stung by bad press and government investigations, Facebook is investing $7.5 million in a partnership with three universities - UC Berkeley, Cornell and Maryland - to develop new methods to improve detection of fake content, fake news and misinformation campaigns. At Berkeley, the work will be led by EECS professors Hany Farid and Alexei Efros.

Graphic of health tools on portable device

Protecting health data privacy

05/01/19 — Artificial intelligence can identify individuals by correlating step data from activity trackers and smartphones with demographic data.

Dawn Song speaking at the EmTech Digital conference

How malevolent machine learning could derail AI

03/26/19 MIT Technology Review — Berkeley Engineering AI security expert Dawn Song warns that “adversarial machine learning” could be used to reverse-engineer systems - including those used in defense.

Activism 2.0: Coding against sex trafficking

01/30/19 — UC Berkeley's ongoing video series on the intersection of social activism and technology profiles recent CS doctoral graduate Rebecca Sorla Portnoff, who uses her computer security know-how to help catch sex traffickers.

AI could identify you and your health history from your step tracker

01/29/19 USA Today — Manufacturers say data stripped of identifying information is no privacy risk. But UC Berkeley's Anil Aswani and UCSF's Yoshimi Fukuoka found that artificial intelligence can overcome that. Time to update health privacy laws.

Artificial intelligence opens health data privacy to attack

12/21/18 — Current privacy laws and regulations are nowhere near sufficient to keep an individual's health data private in the face of advances in artificial intelligence, according to a new study from IEOR professor Anil Aswani and his team.

Doctored photo of shark on a Houston highway

Fighting fake news

11/14/18 — Berkeley students have created SurfSafe, a machine learning tool that identifies when an online photo has been doctored.

Berkeley researchers to help develop trustworthy machine learning systems

10/24/18 — Berkeley engineers, led by computer sciences professor Dawn Song, are part of the new Center for Trustworthy Machine Learning funded by the National Science Foundation. The NSF center, led by Pennsylvania State University and announced today, will focus on developing secure systems in the era of machine learning models. The center will receive $10 million over five years.

The latest course catalog trend? Blockchain 101

09/19/18 Wired — From a course teaching students to think like blockchain entrepreneurs to "collider sprints" at the Sutardja Center's Blockchain Lab, Berkeley is at the forefront of universities incorporating this multidisciplinary technology into the curriculum.

Calling out fake photos on the web

08/21/18 Wired — Fake photos are the bane of internet junkies. SurfSafe, a browser plugin from RoBhat Labs (computer science undergrads Ash Bhat and Rohan Phadte), can warn users that they're viewing a Photoshopped fake in real time - like an antivirus for photos.

Map plotting location history of Google account user

‘Location history’ off? Google’s still tracking you

08/17/18 AP News — Computer science graduate student K. Shankari tipped the Associated Press off to the persistence of Google's movement tracking, even for users who explicitly tell the company not to do so.

Bethany Lyles Goldblum inside Berkeley Lab

Q+A on nuclear nonproliferation

06/01/18 — Bethany Lyles Goldblum is working on new technologies to aid in the detection of nuclear threats.

Illustration of person reading a book with digital assistant in foreground

Alexa and Siri can hear this hidden command. You can’t

05/10/18 New York Times — EECS Ph.D. candidate Nicholas Carlini and other Berkeley cybersecurity researchers have been embedding commands into music and spoken text that human listeners can't hear but smart devices can. Carlini hopes to secure AI systems against attacks that he assumes "malicious people" are already working on.

Archer team members Anjali Banerjee, Tyler Heintz and Alice Ma at Caffe Strada

Scarred by attack, student entrepreneurs fight global terror

04/30/18 — In the wake of the 2016 terrorist attack in Nice, France, that claimed the life of a classmate, Berkeley students in the European Innovation Academy technology entrepreneurship program have embarked on a new mission - fighting global terrorism through startups that are gaining traction far beyond campus.

Q&A with Barbara Simons

03/16/18 — Barbara Simons, a founding member of Women in Computer Science and Engineering (which is celebrating its 40th anniversary), has been sounding the alarm about the potential pitfalls of internet and electronic voting for more than a decade.

Why AI researchers should be more paranoid

02/23/18 Wired — A new report highlights risks of artificial intelligence, such as malicious self-driving cars or assassin robots. EECS professor Ion Stoica, who recently surveyed technical challenges in AI, said Berkeley is already trying to expose students flocking to the field to concerns over AI safety and security.

EECS graduate student Noah Johnson and professor Dawn Song

Security for data analytics: Handling the two-edged sword

02/16/18 Berkeley Research — Data protocols meant to ensure privacy can end up blocking valuable, even life-saving analysis of that data. Now EECS researchers are trying a new approach that allows organizations to follow tight data security and privacy policies while enabling flexible data analysis.

Students in a cryptocurrency course at New York University

Cryptocurrencies come to campus

02/09/18 New York Times — UC Berkeley is one of several institutions rushing to add courses about Bitcoin, blockchain and other elements of virtual currency to the curriculum. EECS professor Dawn Song is co-teaching one such course to a standing-room-only audience of engineering, business and law students.

At Berkeley, ‘ethical hackers’ learn to wage cyberwar

11/29/17 New Yorker — EECS professor Doug Tygar is teaching a new generation of cyber security researchers to prevent cyberwar attacks by forensically examining them, and even sometimes mounting "ethical hacking" schemes of their own.