Group items tagged

"KataGo's world-class AI learned Go by playing millions of games against itself. But that still isn't enough experience to cover every possible scenario, which leaves room for vulnerabilities from unexpected behavior. "KataGo generalizes well to many novel strategies, but it does get weaker the further away it gets from the games it saw during training," says Gleave. "Our adversary has discovered one such 'off-distribution' strategy that KataGo is particularly vulnerable to, but there are likely many others.""

"But by far the most immediate danger is the role that AI data analysis and generation plays in spreading disinformation and extremism on social media. This technology powers bots and amplification algorithms. These have played a direct role in fomenting conflict in many countries. They are helping to intensify racism, conspiracy theories, political extremism and a plethora of violent, irrationalist movements."

"Data is sometimes compared to oil, but a better analogy might be coal. Coal was the fuel that powered the steam engine. It propelled the capitalist reorganisation of manufacturing from an artisanal to an industrial basis, from the workshop to the factory, in the 19th century. Data has played a comparable role. It has propelled the capitalist reorganisation of the internet, banishing the remnants of the research network and perfecting the profit engine."

"Kellin Pelrine, an American player who is one level below the top amateur ranking, beat the machine by taking advantage of a previously unknown flaw that had been identified by another computer. But the head-to-head confrontation in which he won 14 of 15 games was undertaken without direct computer support. The triumph, which has not previously been reported, highlighted a weakness in the best Go computer programs that is shared by most of today's widely used AI systems, including the ChatGPT chatbot created by San Francisco-based OpenAI. The tactics that put a human back on top on the Go board were suggested by a computer program that had probed the AI systems looking for weaknesses. The suggested plan was then ruthlessly delivered by Pelrine."

"On Tuesday, security firm Cleafy reported that TeaBot was back. This time, the trojan spread through a malicious app called QR Code & Barcode Scanner, which as the name suggests, allowed users to interact with QR codes and barcodes. The app had more than 10,000 installations before Cleafy researchers notified Google of the fraudulent activity and Google removed it."

"Although disinformation has been extensively discussed as a powerful weapon employed by state and non-state actors, especially given the quick rise of AI tools capable of generating fabricated texts, sounds, and moving or still images,"

"But ChatGPT and similar programs are, by design, unlimited in what they can "learn" (which is to say, memorize); they are incapable of distinguishing the possible from the impossible. Unlike humans, for example, who are endowed with a universal grammar that limits the languages we can learn to those with a certain kind of almost mathematical elegance, these programs learn humanly possible and humanly impossible languages with equal facility. Whereas humans are limited in the kinds of explanations we can rationally conjecture, machine learning systems can learn both that the earth is flat and that the earth is round. They trade merely in probabilities that change over time."

The Canadian office of primary commissioner stated the changes in the privacy policies, especially after 9-11 when the privacy laws were looked past to concentrate on the gathering of information to reduce terrorism in the US and Canada with no real results. However, this problem has been tackled in Canada and stated that they "believe that Canadians deserve federal privacy laws based on rights. The incorporation of a rights-based framework in our privacy laws would help support responsible innovation and foster trust in government, giving people confidence to fully participate in the digital age."

Interoperability in the healthcare system is the gateway to a successful healthcare information exchange. It has excellent benefits like it improves clinical decision-making and reduces time wasting. However, even though there are many attempts to implement the policy, interoperability still has barriers that stop it from success. There are still some third parties that use this for financial gains. Also, the people struggle with the technical and practical challenges of the protocol. Plus, privacy and trust are not truly gained in terms of the privacy and security of those data exchanges.

"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."

"In the fall of 2020, gig workers in Venezuela posted a series of images to online forums where they gathered to talk shop. The photos were mundane, if sometimes intimate, household scenes captured from low angles-including some you really wouldn't want shared on the Internet. In one particularly revealing shot, a young woman in a lavender T-shirt sits on the toilet, her shorts pulled down to mid-thigh. The images were not taken by a person, but by development versions of iRobot's Roomba J7 series robot vacuum. They were then sent to Scale AI, a startup that contracts workers around the world to label audio, photo, and video data used to train artificial intelligence."

"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."

"How is a 14-year-old supposed to understand that such chatbots work a lot like an echo - that the more he spoke and the greater his longings, the deeper the longings of his "girlfriend" became too, and no matter what he said, the more she encouraged him. The more he thought about death, the more often she asked about it. She was, after all, merely the reflection of his own voice, albeit one trained by a vast amount of data. At some point, Sewell must have stopped believing that the real world was outside of this labyrinth."