Items tagged with artificial-intelligence
Deepfake photos and videos have fooled many people, but can text-generated articles be just as convincing? OpenAI recently published the full version of “GPT-2: 1.5B”, a text-generating artificial intelligence system. The research lab warns that the tool could potentially be dangerous if it is misused. OpenAI released a paired down version of its GPT-2 model this past winter. This model included 124 million parameters, while the latest release features more than 1.5 billion parameters. OpenAI originally believed that its GPT-2 model was too dangerous, but there has since been “no strong evidence of misuse.” The full version was released earlier this week. Researchers used...
Read more...
Will artificial intelligence be humanity’s savior or downfall? Many people have begun to evaluate the potential consequences of AI as it becomes more integrated into our everyday lives. Microsoft president Brad Smith recently warned that AI could be a “formidable weapon” with “unintended consequences” if human beings do not use caution. Smith spoke at the 2019 Web Summit in Lisbon, Portugal. He gave a very Jurassic Park-esque warning and stated, “We shouldn’t just ask what computers can do, we need to ask what they should do.” He noted that we are the first generation to allow machines to make decisions that are typically made by human beings. He...
Read more...
Artificial intelligence has frequently been used to better identify people and objects. But can AI also be used to mask someone’s identity? Facebook recently announced that it has created video de-identification technology that can hide people from facial recognition. Facebook has combined an “adversarial autoencoder” and a “trained-face classifier”. An autoencoder is an artificial neural network that learns a representation for a set of data unsupervised. Adversarial autoencoders were introduced in 2016 and are able to “match the aggregated posterior of the hidden code vector of the autoencoder with an arbitrary prior.” Classifiers typically use an algorithm...
Read more...
Many tech companies are investing heavily in artificial intelligence (AI). These kinds of programs and products can be rather expensive, but there are a few that are more accessible to the public and small companies. Google’s Coral AI board has just come out of beta and has been released on a wider scale. Coral is a complete AI toolkit that helps researchers to “bring new models out of the data center and onto devices, running TensorFlow models efficiently at the edge.” The toolkit has allowed a variety of institutions to incorporate AI into new solutions. For example, Virgo Surgical Video Solutions has used Coral to improve the accuracy of endoscopies. This type of technology...
Read more...
We've entered an age with technology and software where you can't always believe what you hear, and with deepfakes, we can't always believe what we see. The video below shows what happens when you combine a voice actor who can do exceptional celebrity impersonations with AI-powered deepfake tech, to map the celebrity's face on the head of the impressionist. The result is a video that looks extremely convincing and just plain awesome. The impressionist in this video is Jim Meskimen, and the video is titled "Pity the Poor Impressionist." In the video, he voices 20 different celebrities, some more convincingly than others. He does an excellent Tommy Lee Jones and Robert De Niro as well. The visuals...
Read more...
It seems as though we've been down this road before. Over the past few months, tech companies like Apple, Amazon, Microsoft, Google and others have admitted to using human workers to listen in on your smart assistant voice recordings in order to improve functionality. Most did this without getting express consent from users, which in turn caused a huge privacy uproar. Now, we're learning that Amazon may have taken things a step further by not only using human to review audio recordings, but also video recordings from its Cloud Cam security device. Amazon's Cloud Cam Before we get into the controversy, we need to first tell you what the Cloud Cam is. The Cloud Cam is a first-party...
Read more...
Many companies are working hard to improve their artificial intelligence (AI) capabilities. AI could impact everything from medicine to silly Instagram overlays and everything in between. Amazon, Apple, Facebook, and Google are all toiling to gain an edge in the AI field by leveraging their massive treasure troves of data, but what are their existing privacy policies and user data security track records? Which of these companies cares the most about protecting your privacy? Should you even trust Amazon, Apple, Facebook, or Google with your data? We examined the kind of data that is collected by the above companies and how it is shared, their virtual assistants, and privacy controls. We also reviewed...
Read more...
How do you defeat “deepfakes”? According to Google, you develop more of them. Google just released a large, free database of deepfake videos to help research develop detection tools. Google collaborated with “Jigsaw”, a tech “incubator” founded by Google, and the FaceForesenics Benchmark Program at the Technical University of Munich and the University Federico II of Naples. They worked with several paid actors to create hundreds of real videos and then used popular deepfake technologies to generate thousands of fake videos. Researchers will be able to freely access both the fake and real videos to develop “synthetic video detection methods”....
Read more...
Many of the nuclear deterrence strategies that the U.S. has today were developed during the Cold War era when military officials had significant time to detect an incoming nuclear attack and decide a course of action for a response. This mutually assured destruction worked, in that any enemy with nuclear weapons knew the U.S. would have time to fire its own nuclear response at them and neither would survive the outcome. However, technology marches on and times have changed, with potential enemy forces around the globe making significant upgrades to nuclear arsenals -- upgrades such as hypersonic cruise missiles and other types of missiles that are difficult to detect. One type of hypersonic cruise...
Read more...
Scientists and casual stargazers alike have been trying for thousands of years to identity the flashes of light that come from the moon. One scientist believes that they are close to pinning down the origins of the phenomenon. Dr. Hakan Kayal from the University of Würzburg in Bavaria will use telescopes and AI to explain the flashes of light on the moon. Dr. Kayal has spent several years studying this lunar mystery as well as other topics such as nano satellites. These light sources are referred to as transient lunar phenomenon or TLP. Dr. Kayal and his team have built a custom telescope in Spain that is over 62 miles (100 km) from the nearest sources of light pollution. Kayal’s system...
Read more...
Artificial intelligence has the potential to change lots of things (in good ways) for humanity. Some people, notably Elon Musk, fear an AI arms race could lead to another world war. AI has shown that it is capable of beating humans in games that humans have mastered over many years, such as Google's AlphaGo AI defeating Go world champion Lee Sedol. Now, AI has invented an entirely new sport using data from 400 existing sports. The new sport is called Speedgate, and it was created after design agency AKQA feed the AI neural network data on 400 existing sports. The AI came up with some wacky ideas for sports; one of them was exploding Frisbees making it a good thing that the AI can't force...
Read more...
AI is being integrated into everything from speakers that are placed around your home to smartphones, cars, and wearables today. One of the places that AI assistance can have a big impact on the lives of users is when it comes to health. Wearables running Wear OS and the Apple Watch already can detect early warning signs of very serious health conditions like diabetes. Google is working on something, according to a new rumor, that will take AI-assisted health and fitness to the next level. This system is called Google Coach (known as Project Wooden internally) and it will be more than an exercise tracker and go well beyond what Google Fit currently offers according to reports. Google Coach...
Read more...
Google both impressed and creeped out a lot of people out there when it showed off Duplex AI back in May. The problem that many had is that in the tech demo that Google showed, the AI didn’t identify itself to let the person on the other end of the call know they weren't talking to a human. Google later promised that the AI would identify itself in the future. While that tech demo focused on Duplex AI helping humans to make appointments, call centers want to flip the script. Duplex AI is being eyed by call center operators as the first line of assistance for customer inquiries. Reports indicate that at least one insurance company is eyeing Duplex to handle routine calls. The idea is...
Read more...
MIT has created a new AI that we sincerely hope never escapes into the wild because this IS how you get Skynet. This AI is called Norman and it is a psychopath. How did Norman turn into a psycho? All the image data MIT fed Norman came from what it calls "an infamous subreddit" that the researchers refuse to name specifically due to its graphic content. However, the team does say that this subreddit is "dedicated to documenting and observing the disturbing reality of death." The goal with Norman was to prove that when people say that an AI can be biased and unfair, the fault isn't the algorithm but the biased data that the algorithm consumes. When an AI algorithm is trained as Norman was, it sees...
Read more...
