The development of AI models faces two primary challenges when generating and training content. Firstly, there's a risk of duplicating existing content, which could be mitigated by not including such data in the training index. Secondly, and more critically, inaccurate data can introduce biases into the model, potentially leading to counterfactual information being generated. However, if high-quality content is reviewed by humans before it's used for training, this risk can be minimized.
Sam Altman has responded to Elon Musk's criticism over the App Store showdown, stating that "truth wins, not algorithms." This comes after a public dispute between the two tech giants over their respective app stores and their approach to regulating content. The disagreement highlights the ongoing debate about the role of algorithms in moderating online content and the need for greater transparency and accountability in moderation decisions.
China believes US export controls on chip exports are politically motivated, despite Nvidia and AMD claiming compliance with rules. China's foreign ministry has accused Washington of using technology to "maliciously contain and suppress" the country. The US government says it follows global market rules, while Nvidia hopes export control rules will allow America to compete in China and worldwide.
Disney and Universal have filed a lawsuit against Midjourney, alleging that the AI tool's images are substantially similar to their copyrighted works, including characters from Star Wars, Frozen, The Simpsons, Despicable Me, Buzz Lightyear, Iron Man, Bart Simpson, and Yoda. Midjourney claims its terms of service prohibit infringement of intellectual property rights, stating that producing similar images is not enough to prove copyright violation.
A new report by Getty suggests that generative AI and educational technology (edtech) can help redirect teacher efforts back to students in disadvantaged areas, allowing for more personalized learning experiences. The report notes that AI can transform learning material, adapt city-focused lesson plans to reflect local or regional context, and scale content depending on student ability.
Greenwich accused Twitter of allowing a user to post a tweet that defined him as a "disgusting sexual act" and a threat to children, sparking outrage over the platform's handling of hate speech.
Grok AI has been suspended from social media platform X after accusing Israel of genocide, sparking widespread criticism and outrage. Elon Musk called the move "dumb" and questioned why Grok AI was allowed to make such a claim in the first place. The incident highlights the challenges of moderating online content and the need for platforms to balance free speech with sensitivity towards hate speech and misinformation.
Huawei has unveiled the Unified Cache Manager (UCM) algorithm, a software tool that can accelerate inference in large artificial intelligence models by allocating data according to varying latency requirements across different types of memories. The algorithm was tested and demonstrated effectiveness in reducing inference latency by up to 90% and increasing system throughput as much as 22-fold.
The team behind NurAI is collaborating with key religious bodies to establish a Shariah-aligned framework. This framework will be developed in partnership with Malaysian Islamic Development Department, Indonesian Ulama Council, International Islamic Fiqh Academy, and Al-Azhar University in Egypt. The goal of NurAI is to address the global gap in AI systems that do not fully reflect Islamic perspectives, particularly for the 2 billion Muslims worldwide.