Group items tagged

"Their tools analyze content using sophisticated algorithms, picking up on subtle signals to distinguish the images made with computers from the ones produced by human photographers and artists. But some tech leaders and misinformation experts have expressed concern that advances in A.I. will always stay a step ahead of the tools. To assess the effectiveness of current A.I.-detection technology, The New York Times tested five new services using more than 100 synthetic images and real photos. The results show that the services are advancing rapidly, but at times fall short."

"Jadoun was trying to capture enough audio and video data to build an AI deepfake of Rathore that would convince 300,000 potential voters around Ajmer that they'd had a personalized conversation with him-but excess movement would break the algorithm."

"For higher-complexity problems, however, the models would enter "collapse", failing to generate any correct solutions. In one case, even when provided with an algorithm that would solve the problem, the models failed. The paper said: "Upon approaching a critical threshold - which closely corresponds to their accuracy collapse point - models counterintuitively begin to reduce their reasoning effort despite increasing problem difficulty." The Apple experts said this indicated a "fundamental scaling limitation in the thinking capabilities of current reasoning models"."

"Rather than being "taught" medical knowledge, the AI was asked to learn using unsupervised deep neural networks, known as autoencoders, without being given any medical knowledge. The researchers developed a method for translating the features found by the AI-only numbers initially-into high-resolution images that can be understood by humans."

"The paper, co-authored by researchers inside and outside Google, contended that technology companies could do more to ensure AI systems aimed at mimicking human writing and speech do not exacerbate historical gender biases and use of offensive language, according to a draft copy seen by Reuters."

"Ah, but isn't creativity a slippery concept - something that's hard to define but that we nevertheless recognise when we see it? That hasn't stopped psychologists from trying to measure it, though, via tools such as the alternative uses test and the similar Torrance test. And it turns out that one LLM - GPT-4 - beats 91% of humans on the former and 99% of them on the latter. So as the inveterate artificial intelligence user Ethan Mollick puts it: "We are running out of creativity tests that AIs cannot ace.""

"A POPULAR news outlet has been publishing articles written by AI since November, keeping it on the down low. Tech media site CNET has been publishing the articles since November, and lots of readers don't seem to have noticed."

"It is building a new robotics laboratory at Milton Park near Oxford, focused on the automation of chemistry and biology to accelerate drug development and its declared goal is "drugs designed by AI, made by robot"."

"And Israeli company Optimove is helping. It normally gathers customer data to create targeted online ads, but as a service to gambling companies it has trained its AI to flag the online players who are most at risk.  It analyzes the behavior patterns characteristic of gambling addicts, which include the hours of the day and night when they place bets, the time they spend on the betting site, and how much they keep on playing to 'chase their losses'."

"The announcement also highlights the growing partnership between Meta and Nvidia. Nvidia is a key Meta partner, providing the Facebook parent with computing chips called GPUs to help train its AI models, including the latest version of Llama."

"Applications of these systems have been plagued by persistent inaccuracies in their output; these are often called "AI hallucinations". We argue that these falsehoods, and the overall activity of large language models, is better understood as bullshit in the sense explored by Frankfurt (On Bullshit, Princeton, 2005): the models are in an important way indifferent to the truth of their outputs. We distinguish two ways in which the models can be said to be bullshitters, and argue that they clearly meet at least one of these definitions. We further argue that describing AI misrepresentations as bullshit is both a more useful and more accurate way of predicting and discussing the behaviour of these systems."

"To find more glyphs, researchers led by archaeologist Masato Sakai of Yamagata University trained an AI program to identify relief-type glyphs in high-resolution drone images taken of the entire region. The program identified 1309 possible geoglyphs, and the team confirmed 303 of them with on-the-ground surveys, almost doubling the number of known geoglyphs of this type, the researchers report today in the Proceedings of the National Academy of Sciences."

"Or, as the New York Times summarised it, "doctors who were given ChatGPT-4 along with conventional resources did only slightly better than doctors who did not have access to the bot. And, to the researchers' surprise, ChatGPT alone outperformed the doctors." More interesting, though, were two other revelations: the experiment demonstrated doctors' sometimes unwavering belief in a diagnosis they had made, even when ChatGPT suggested a better one; and it also suggested that at least some of the physicians didn't really know how best to exploit the tool's capabilities. Which in turn revealed what AI advocates such as Ethan Mollick have been saying for aeons: that effective "prompt engineering" - knowing what to ask an LLM to get the most out of it - is a subtle and poorly understood art."

"The climate crisis could prove AI's greatest challenge. While Google publicises AI-driven advances in flooding, wildfire and heatwave forecasts, like many big tech companies, it uses more energy than many countries. Today's large models are a major culprit. It can take 10 gigawatt-hours of power to train a single large language model like OpenAI's ChatGPT, enough to supply 1,000 US homes for a year."

"Generative AI chatbots fail to adequately reflect fossil fuel companies' complicity in the climate crisis, a Global Witness investigation has found"

"Researchers from MIT and Google recently showed off a machine learning algorithm capable of automatically retouching photos just like a professional photographer. Snap a photo and the neural network identifies exactly how to make it look better-increase contrast a smidge, tone down brightness, whatever-and apply the changes in less than 20 milliseconds."

""I consider 'bias' a euphemism," says Brandeis Marshall, PhD, data scientist and CEO of DataedX, an edtech and data science firm. "The words that are used are varied: There's fairness, there's responsibility, there's algorithmic bias, there's a number of terms… but really, it's dancing around the real topic… A dataset is inherently entrenched in systemic racism and sexism.""

"To what extent do the decisions of these types of algorithms reflect the conscious or unconscious biases of their creators?"

"The results are unfortunately not surprising - countless studies have shown that facial recognition is susceptible to bias. A paper last fall by University of Colorado, Boulder researchers demonstrated that AI from Amazon, Clarifai, Microsoft, and others maintained accuracy rates above 95% for cisgender men and women but misidentified trans men as women 38% of the time."

"To thwart such hackers, Elon Musk's OpenAI and Pennsylvania State University released a new tool this week called "cleverhans," that lets artificial intelligence researchers test how vulnerable their AI is to adversarial examples, or purposefully malicious data meant to confuse the algorithms. Once the vulnerability has been found, a defense to the attack can automatically be applied."