2026
2026
October
Cranor weighs in on privacy concerns regarding smart glasses
Communications of the ACM
As smart glasses become a more common piece of wearable tech, concerns are being raised about their data collection and protection practices. CyLab Director Lorrie Cranor says the solution is to integrate privacy from the start in the design of these AI-integrated wearable devices.
Sekar sheds light on AI’s relationship with home smart devices
Scientific American
ECE/CyLab’s Vyas Sekar was quoted on the potential consequences AI might have on homes full of smart devices in response to Anthropic’s recent blog post revealing AI’s unique capabilities as computer worms, programs that leave copies of themselves behind. Sekar’s insight reveals that more people will be able to launch malicious programs or malware using AI.
Mireshghallah’s AI study featured for revealing adversarial delegation
Inc.
EPP/CyLab’s Niloofar Mireshghallah co-authored a study alongside Aman Priyanshu, Supriti Vijay, and Brian Jabarian that revealed how AI assistants have provided more expensive options to wealthier users, even when they ask for the cheapest options.
September
Cranor discusses the future of login privacy and functionality
The Atlantic
CyLab Director Lorrie Cranor was interviewed for a feature story about the inconvenience created when browsers and operating systems compete with third-party password managers to generate, store, and autofill your credentials. She discusses how passkeys work and often frustrate users trying to log in.
July
Sekar quoted on need for federal AI guardrails
Channel News Asia
ECE’s Vyas Sekar spoke to Channel News Asia about the need for increased regulations around AI development in the U.S. This follows a petition calling for such guardrails signed by over 1,000 employees at leading AI companies.
June
Sekar quoted on social media fingerprint hacking worries
CBS News
ECE’s Vyas Sekar was quoted in a CBS News article responding to viral concerns that a hacker could lift a person’s fingerprints from a photo posted on social media. He said that the risk was possible in theory, but that it would be difficult for hackers to actually target someone this way.