Toggle light / dark theme

OpenAI Said Its Code Was Risky. Two Grads Re-Created It Anyway

In February, an artificial intelligence lab cofounded by Elon Musk informed the world that its latest breakthrough was too risky to release to the public. OpenAI claimed it had made language software so fluent at generating text that it might be adapted to crank out fake news or spam.

On Thursday, two recent master’s graduates in computer science released what they say is a re-creation of OpenAI’s withheld software onto the internet for anyone to download and use.

Aaron Gokaslan, 23, and Vanya Cohen, 24, say they aren’t out to cause havoc and don’t believe such software poses much risk to society yet. The pair say their release was intended to show that you don’t have to be an elite lab rich in dollars and PhDs to create this kind of software: They used an estimated $50,000 worth of free cloud computing from Google, which hands out credits to academic institutions. And they argue that setting their creation free can help others explore and prepare for future advances—good or bad.

Researchers Created AI That Hides Your Emotions From Other AI

Humans can communicate a range of nonverbal emotions, from terrified shrieks to exasperated groans. Voice inflections and cues can communicate subtle feelings, from ecstasy to agony, arousal and disgust. Even when simply speaking, the human voice is stuffed with meaning, and a lot of potential value if you’re a company collecting personal data.

Now, researchers at the Imperial College London have used AI to mask the emotional cues in users’ voices when they’re speaking to internet-connected voice assistants. The idea is to put a “layer” between the user and the cloud their data is uploaded to by automatically converting emotional speech into “normal” speech. They recently published their paper “Emotionless: Privacy-Preserving Speech Analysis for Voice Assistants” on the arXiv preprint server.

Our voices can reveal our confidence and stress levels, physical condition, age, gender, and personal traits. This isn’t lost on smart speaker makers, and companies such as Amazon are always working to improve the emotion-detecting abilities of AI.

Physicists’ study demonstrates silicon’s energy-harvesting power

A University of Texas at Dallas physicist has teamed with Texas Instruments Inc. to design a better way for electronics to convert waste heat into reusable energy.

The collaborative project demonstrated that silicon’s ability to harvest energy from heat can be greatly increased while remaining mass-producible.

Dr. Mark Lee, professor and head of the Department of Physics in the School of Natural Sciences and Mathematics, is the corresponding author of a study published July 15 in Nature Electronics that describes the results. The findings could greatly influence how circuits are cooled in electronics, as well as provide a method of powering the sensors used in the growing “internet of things.”

Employees connect nuclear plant to the internet so they can mine cryptocurrency

Ukrainian authorities are investigating a potential security breach at a local nuclear power plant after employees connected parts of its internal network to the internet so they could mine cryptocurrency.

The investigation is being led by the Ukrainian Secret Service (SBU), who is looking at the incident as a potential breach of state secrets due to the classification of nuclear power plants as critical infrastructure.

Investigators are examining if attackers might have used the mining rigs as a pivot point to enter the nuclear power plant’s network and retrieve information from its systems, such as data about the plant’s physical defenses and protections.

Practical anonymous communication protocol developed for quantum networks

The ability to securely transmit information over the internet is extremely important, but most of the time, eavesdroppers can still generally determine who the sender and receiver are. In some highly confidential situations, it is important that the sender’s and receiver’s identities remain anonymous.

Over the past couple of decades, researchers have been developing protocols for anonymously transmitting messages over classical networks, but similar protocols for are still in much earlier stages of development. The anonymity methods that have been proposed for quantum networks so far face challenges such as implementation difficulties or require that strong assumptions be made about the resources, making them impractical for use in the .

In a new paper, Anupama Unnikrishnan, Ian MacFarlane, Richard Yi, Eleni Diamanti, Damian Markham, and Iordanis Kerenidis, from the University of Oxford, MIT, Sorbonne University, the University of Paris and CNRS, have proposed the first practical for anonymous communication in quantum networks.

A Tech Expert Says Google Chrome Has Become Spy Software

You open your browser to look at the Web. Do you know who is looking back at you?

Over a recent week of Web surfing, I peered under the hood of Google Chrome and found it brought along a few thousand friends. Shopping, news and even government sites quietly tagged my browser to let ad and data companies ride shotgun while I clicked around the Web.

This was made possible by the Web’s biggest snoop of all: Google. Seen from the inside, its Chrome browser looks a lot like surveillance software.

Security Researchers Show Hackers Can Take Over Some Cameras With Ransomware

A group of security researchers has found that some cameras are vulnerable to ransomware attacks via Wi-Fi or hijacked computers (using the USB connection) that can render a camera and its photos entirely usage until a sum of money is paid to the hackers.

Check Point Software Technologies has shown (as seen in the video above) that Canon DSLRs (it’s not clear if other cameras are also affected) are susceptible to ransomware attacks, an increasingly common exploitation in which a hacker disables a device until the owner pays a ransom to regain control. As more and more electronics gain internet connectivity, these attacks are becoming more common, particularly since manufacturers typically don’t put the same sort of effort into fortifying the network security of small electronics as they do into computers. Canon released a statement in which they offered some barebones tips to avoid being susceptible to these attacks and noted that there have been no known instances of such an attack occurring. They did not indicate whether they plan to fix the vulnerabilities through firmware updates, so it probably doesn’t hurt to be a bit more careful to protect your photos.

This Hydroponic Farm Is Run Entirely By Robots

Iron Ox has just opened its first fully automated farm in San Carlos, California. The company claims that their hydroponic system can produce 30 times the yield per acre of land comparing to traditional farms, while using 90% less water.

► Subscribe for more tech & culture videos: http://on.mash.to/subscribe
MORE FROM MASHABLE
‌• Video — https://youtu.be/6ElleCub6uM
‌• Video — https://youtu.be/jAu1ZsTCA64
‌• Video — https://youtu.be/W8T6fYt9-tA

Best of playlist: https://on.mash.to/BestOf

MASHABLE ACROSS THE WEB
Mashable.com: http://on.mash.to/1hCcRpl
Facebook: http://on.mash.to/2lyOwmZ
Twitter: http://on.mash.to/1Udp1kz
Instagram: http://on.mash.to/1U6D40z
Mashable is for superfans. We’re not for the casually curious. Obsess with us.

/* */