Group items tagged

"Let's say you have an IoT device that is well protected with functions that can be accessed through a well-defined API; even if you can control the device through the API, you probably can't do too much harm. Firmware, the programming inside the device has logical constraints that don't allow you, for example, to close garage doors while someone is in the way of them or overheat a device so that it combusts.  We used to trust that hardware, such as a common kitchen appliance, could be trusted and could not be easily altered without physically dismounting the device. But with today's "smart" appliances, this is no longer the case."

"The use of AI in military applications - such as small anti-personnel weapons - is of particular concern, he said. "Those are the ones that are very easily scalable, meaning you could put a million of them in a single truck and you could open the back and off they go and wipe out a whole city," said Russell."

"At the Seco Innovation Laboratory in Fagersta, Sweden, R&D technician Jan Gravningsbråten, could easily be mistaken for a mad scientist, surrounded by blinking red lasers, trays of inserts under observation, diagnostic equipment and countless other measuring devices, all in aid of his ongoing work with data matrix codes, also known as QR codes."

""The social engineering type of attack does not tend to scale [up] easily given the time and effort required to succeed, and therefore is more often than not used by individuals rather than the 'call centre' approach of criminal enterprises," Goddard says. "The trigger to target an individual could be targeted, or opportunistic such as overhearing a conversation or getting access to sensitive or exploitable information like a picture or bank statement.""

"PhotoGuard works by altering select pixels in an image such that they will disrupt an AI's ability to understand what the image is. Those "perturbations," as the research team refers to them, are invisible to the human eye but easily readable by machines. The "encoder" attack method of introducing these artifacts targets the algorithmic model's latent representation of the target image - the complex mathematics that describes the position and color of every pixel in an image - essentially preventing the AI from understanding what it is looking at."

"Musicians are therefore worried - about being made to perform material they otherwise wouldn't, or being usurped by a fantasy. "I can't help but think that I can be easily replaced," says Flora Rose, a singer-songwriter on TikTok. "I'm spending months crafting my debut EP, [and meanwhile] people can make tracks in one click." When it comes to the arts, AI tends to provoke horror or ridicule - as when an AI photograph won a major photography competition, or when ChatGPT declared young adult weepie The Fault in Our Stars "one of the best books of all time". In February, the lawyer behind a lawsuit on behalf of visual artists whose work was being used to generate AI art called any generative image "an infringing derivative work"."

"We submitted the grant on time. The next day, while speaking to a friend, I told him, "This week, I wrote my first ChatGPT grant." He replied that he had been doing it for months and that many other scientists are doing the same. A 2023 Nature survey of 1,600 researchers found that more than 25% use AI to help them write manuscripts and that more than 15% use the technology to help them write grant proposals. Some people might see the use of ChatGPT in writing grant proposals as cheating, but it actually highlights a much bigger problem: what is the point of asking scientists to write documents that can be easily created with AI? What value are we adding? Perhaps it is time for funding bodies to rethink their application processes."

"As if robocalling wasn't already enough of a problem, the advent of easily accessible, realistic AI-powered writing and synthetic voice could supercharge the practice. The FCC aims to preempt this by looking into how generated robocalls might fit under existing consumer protections. A Notice of Inquiry has been proposed by Chairwoman Jessica Rosenworcel to be voted on at the agency's next meeting. If the vote succeeds (as it is almost certain to), the FCC would formally look into how the Telephone Consumer Protection Act empowers them to act against scammers and spammers using AI technology. But Rosenworcel was also careful to acknowledge that AI represents a potentially powerful tool for accessibility and responsiveness in phone-based interactions. "While we are aware of the challenges AI can present, there is also significant potential to use this technology to benefit communications networks and their customers-including in the fight against junk robocalls and robotexts. We need to address these opportunities and risks thoughtfully, and the effort we are launching today will help us gain more insight on both fronts," she said in a statement."

"Researchers put AI normally used to search for helpful drugs into a kind of "bad actor" mode to show how easily it could be abused at a biological arms control conference. All the researchers had to do was tweak their methodology to seek out, rather than weed out toxicity. The AI came up with tens of thousands of new substances, some of which are similar to VX, the most potent nerve agent ever developed. Shaken, they published their findings this month in the journal Nature Machine Intelligence."

""We are seeing deep fakes being used all the time, and the technology is going to allow still images, but ultimately also video images, to be synthesised [more easily] by bad actors," he says. DALL-E has content policy rules in place that prohibit bullying, harassment, the creation of sexual or political content, or creating images of people without their consent. And while Open AI has limited the number of people who can sign up to DALL-E, its lower-grade replica, DALL-E mini, is open access, meaning people can produce anything they want."

This discrepancy has made it relatively easy to connect an anonymous line of data to a specific person: if a private detective is searching for someone in New York City and knows the subject is male, is 30 to 35 years old and has diabetes, the sleuth would not be able to deduce the man's name-but could likely do so quite easily if he or she also knows the target's birthday, number of children, zip code, employer and car model.

"In the future, it may be possible to guard against this kind of photo misuse through technical means. For example, future AI image generators might be required by law to embed invisible watermarks into their outputs so that they can be read later, and people will know they're fakes. But people will need to be able to read the watermarks easily (and be educated on how they work) for that to have any effect. Even so, will it matter if an embarrassing fake photo of a kid shared with an entire school has an invisible watermark? The damage will have already been done."