Sept. 10, 2024, 8:23 p.m. • Neuroscience News • (4 Minute Read)
In a pioneering study, Dr. Patrick Krauss and Dr. Achim Schilling from the Cognitive Computational Neuroscience Group at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) used artificial intelligence to gain major insights into how the brain predicts future events and processes information. By analyzing local field potentials (LFPs), they discovered that the brain remains active in anticipating possible scenarios, even in a resting state. This research may lead to better diagnostic tools and treatments for neurological diseases by improving our understanding of how thoughts and feelings are processed in our minds. The study also reveals the potential for AI to continuously make predictions, even without processing any input, which could have implications for various applications, including vehicle safety. The results show how the fusion of technology and brain research is expanding the boundaries of cognitive processes and brain function, ultimately leading to innovative approaches in medical diagnosis and therapy.
The findings were published in NeuroImage and demonstrate the potential of using unsupervised machine learning to extract meaningful information from single-trial electro-physiological recordings, offering new insights into how information is processed in the cerebral cortex. This research opens up new avenues for understanding the complexities of the human brain, with the potential to improve diagnosis and treatment for a range of neurological diseases.
Sept. 9, 2024, 6:34 a.m. • The Register • (4 Minute Read)
In a rare joint statement, CIA director Bill Burns and UK Secret Intelligence Service (SIS) chief Richard Moore revealed that their agencies have adopted generative AI to combat tech-driven threats from Russia and China. They stated that AI, including generative AI, is being used to improve intelligence activities, protect operations, and analyze vast amounts of data. The use of cloud technologies and collaboration with innovative companies were also emphasized. Moore disclosed that MI6 uses large language models to navigate extremist content and decode criminal vernacular on the internet. The intelligence bosses highlighted the unique challenges posed by technology, especially in the context of the conflict in Ukraine, which demonstrated the combination of various technologies at a rapid pace. They also emphasized efforts to disrupt disinformation campaigns and sabotage by Russia as well as the increasing focus on addressing technology theft and security issues from China.
July 8, 2024, 4 a.m. • MIT News • (6 Minute Read)
MIT researchers have introduced GenSQL, a generative AI system for databases that simplifies the analysis of complex tabular data. This new tool enables users to perform statistical analyses, prediction, anomaly detection, and data generation with just a few keystrokes, without the need to understand the underlying mechanisms. GenSQL integrates a tabular dataset with a generative probabilistic AI model to account for uncertainty and adapt decision-making based on new data. Compared to popular AI-based approaches for data analysis, GenSQL is not only faster but also produces more accurate results. The system's goal is to enable a large set of users to query their data and models without needing intricate technical knowledge. The research was recently presented at the ACM Conference on Programming Language Design and Implementation and is funded by organizations like DARPA and Google. The researchers plan to further develop GenSQL to conduct large-scale modeling of human populations and eventually enable natural language queries.
July 7, 2024, 10 a.m. • South China Morning Post • (1 Minute Read)
Shanghai has released China's first set of guidelines for governing humanoid robots, emphasizing the need to ensure these machines do not pose a threat to human security and effectively safeguard human dignity. The guidelines, unveiled during the World Artificial Intelligence Conference, call for measures such as risk warning procedures, emergency response systems, and ethical and lawful use training for users. The document, developed by a coalition of Shanghai-based industry organizations, also advocates for international cooperation in the humanoid robot sector, recommending the establishment of a global governance framework and an international think tank dedicated to governing these machines. Chinese companies are racing to develop cost-effective humanoid robots, as the country aims for mass production by 2025 and global leadership in the sector by 2027. Tesla's second-generation humanoid robot, Optimus, and other cutting-edge models were showcased at the conference, highlighting China's efforts to catch up with the US in AI and achieve technological self-sufficiency.
July 5, 2024, 2 a.m. • South China Morning Post • (1 Minute Read)
Tesla unveiled its humanoid robot, Optimus, at the 2024 World Artificial Intelligence Conference in Shanghai, sparking interest as one of the few American AI products at the event. Despite being showcased behind glass and without interactive capabilities, Optimus drew attention with its potential to handle various tasks using Tesla's neural network and computer vision technology. The display also featured 17 other robots from Chinese manufacturers, highlighting the increasing development and potential applications of humanoid robots in sectors such as education, entertainment, healthcare, elder care, and manufacturing. However, high production costs remain a barrier to widespread deployment, with prices ranging from $70,000 to $1 million for existing models, although Tesla's Optimus is anticipated to sell for up to $30,000.
July 3, 2024, 7:44 p.m. • The Register • (6 Minute Read)
Cloudflare has introduced a new feature enabling web hosting customers to block AI bots from scraping website content and using it without permission for training machine learning models. This move comes in response to customer concerns about dishonest AI bot visits and the unauthorized usage of website data. While the robots.txt file already offers a means to block bots, Cloudflare's new feature provides a one-click option to block all AI bots. The company's decision reflects the growing unease about AI companies using web content without consent, especially for training AI models. Cloudflare's new offering aims to provide a more robust defense against AI bots, which currently visit around 39 percent of the top one million web properties served by Cloudflare. This new tool is available to all customers and is located in the Security -> Bots menu for a given website, reflecting Cloudflare's commitment to helping content creators maintain control over their content.
June 17, 2024, 9 p.m. • Futurism • (2 Minute Read)
The recently resurfaced video of an online lecture by AI pioneer Geoffrey Hinton has stirred controversy, in which he expresses his support for the idea of a superintelligent AI replacing humanity. Despite his later departure from Google due to concerns about AI misuse, Hinton's statement is surprising. When pressed to explain his position, he humorously suggests that humanity may not be the best form of intelligence and questions the use of terms like "humanist." This perspective contrasts with his earlier assertion that human-level artificial general intelligence could potentially destroy humanity. Given the conflicting nature of his statements, further clarification from Hinton would be valuable in understanding the evolution of his views on AI and its implications for society.
June 9, 2024, 9:15 p.m. • VentureBeat • (5 Minute Read)
In a recent news story titled "The AI paradox: Path to utopia or dystopia?", the debate over the responsible development and deployment of AI has been reignited by recent headlines, such as an AI suggesting people should eat rocks or the creation of 'Miss AI,' the first beauty contest with AI-generated contestants. While there have been concerns about potential risks from AI technology, including its weaponization, encoding of societal biases, and privacy violations, there is no evidence yet that AI on its own is posing harm. The emergence of advanced AI models and the potential achievement of artificial general intelligence (AGI) within five years have raised existential worries, but disagreement exists about whether this progress will continue or lead to a new "AI Winter." Responsible AI development with ethical principles, safety testing, human oversight, and control measures is considered essential to navigating the rapidly evolving AI landscape. This article presents different viewpoints and acknowledges the uncertainty surrounding AI's future impact on society.
June 8, 2024, 11:40 a.m. • Gizmodo • (5 Minute Read)
Paleontologists are expressing frustration over the use of AI to depict prehistoric animals, as the technology's imagery has sparked controversy. The debate revolves around the accuracy and scientific integrity of AI-generated depictions, with concerns that they may diverge from scientific findings and expectations. While AI has potential in the sciences, including in paleontology where it aids in tasks such as analyzing fossils, experts argue that AI-generated paleoart lacks the nuanced and informed approach of human artists. The use of AI in scientific visual communication has raised questions about transparency, ethical considerations, and maintaining the public's trust in scientific accuracy. These developments prompt a call for clear guidelines and ethical standards to govern AI's role in communicating scientific information.
June 5, 2024, 12:16 p.m. • Fortune • (4 Minute Read)
Elon Musk has admitted to diverting Nvidia's latest AI chips originally destined for Tesla to his other companies, such as his xAI startup and social media company, X. This move has stirred ongoing concerns over a conflict of interest, especially given Musk's recent pivot to rebranding Tesla as an AI and robotics company. Musk justified his decision by claiming that Tesla had no immediate use for the chips and that they would have otherwise sat in a warehouse. However, this revelation raises questions about potential breaches of fiduciary responsibility to Tesla shareholders, as well as the competition for resources between Musk's publicly traded and privately owned companies. Moreover, with the looming vote on Musk's record pay package and previous disputes over stock options, shareholders are increasingly wary of potential impacts on Tesla's operations and investments in AI technology.
June 4, 2024, 4 p.m. • TechCrunch • (7 Minute Read)
True Fit, the AI-powered size-and-fit personalization tool, has unveiled its latest offering, "Fit Hub," a tool that utilizes generative AI to revolutionize the online shopping experience. The tool's objective is to address the ongoing struggle of consumers with sizing issues, leading to high return rates for e-commerce purchases. Fit Hub consolidates information from product pages to help shoppers understand an item and make confident sizing decisions, utilizing various generative AI models to process data at faster speeds and greater volumes. This new tool aims to minimize returns and enhance the overall online shopping experience, marking a significant advancement in the e-commerce industry. True Fit is currently conducting beta testing with several brands and plans to make Fit Hub available to all its merchant partners soon. The long-term roadmap includes additional features, such as "Shopper Insights" and "Brand Sizing," demonstrating True Fit's commitment to enhancing the online shopping experience for consumers by leveraging the power of generative AI.
June 4, 2024, 7:01 a.m. • The Register • (4 Minute Read)
Raspberry Pi has unveiled a new addition to its single-board computer, the Hailo-8L AI accelerator. The AI Kit, priced at $70, features an M.2 Hat board with a preinstalled M.2 stick carrying the Hailo-8L accelerator, marking the first time Raspberry Pi has offered a first-party AI accessory. Despite offering 13 TOPS of performance at INT8, the Hailo-8L is competitively priced and boasts low power consumption, making it an attractive option for exploring machine learning. Raspberry Pi CEO Eben Upton emphasized the company's focus on a disaggregated architecture with the Raspberry Pi 5, explaining that offering the AI Kit as a separate accessory allows users to purchase the Raspberry Pi 5 at a lower price and add the Hailo-8L if desired. This development has the potential to reduce reliance on cloud connectivity for AI workloads and is expected to spur a significant increase in AI applications and solutions. Upton hinted at future collaborations between Raspberry Pi and Hailo, suggesting further advancements in AI hardware may be on the horizon.
May 29, 2024, 7:53 a.m. • Eurogamer • (9 Minute Read)
The article "How AI is changing video game development forever" discusses the increasing role of AI in video game development, covering various aspects such as animation, voice acting, and artwork. It also examines the ethical and legal implications of AI in gaming, including the use of AI voice cloning and concerns about copyright infringement. The article presents insights from AI expert Dr. Tommy Thompson, who discusses the definition of AI and its current and potential uses in game development. Thompson highlights both the positive and negative impacts of AI and suggests that the industry may be heading towards an "AI winter" due to the lack of substantial returns on investment in AI technologies. Additionally, the article addresses the shift towards in-house AI tools by game studios and the challenges faced by indie developers in utilizing third-party AI tools. Thompson also emphasizes the importance of human input in game creation and expresses caution about the industry's current hype surrounding AI, suggesting that there is increasing apathy and pessimism towards generative AI. The article offers a comprehensive exploration of the evolving intersection of AI and video game development, presenting a balanced perspective on its potential impact on the industry.
May 28, 2024, 10:52 p.m. • NBC News • (2 Minute Read)
A bilingual brain implant using AI technology has been developed to help a stroke survivor communicate in both Spanish and English. Researchers at the University of California, San Francisco have successfully designed a decoding system that allows the man, known as Pancho, to produce sentences in both languages using his brain activity, which are then displayed on a screen. Dr. Edward Chang and his team utilized an AI method called neural network to train the implant to decode words based on Pancho’s brain activity. The study, published in Nature Biomedical Engineering, offers a promising glimpse into the potential of this technology to restore natural communication for bilingual speakers with paralysis.
May 22, 2024, 7:08 p.m. • Neuroscience News • (4 Minute Read)
In a bold move that pushes the boundaries of medical science, BrainBridge, a pioneering neuroscience and biomedical engineering start-up, unveiled a groundbreaking concept for a “head transplant system.” The company envisions a fully automated, robotic system capable of performing this complex procedure within the next eight years. The potential implications of this technology are immense, offering renewed hope to individuals suffering from terminal illnesses, neurodegenerative diseases, and paralysis. Using advanced robotics, artificial intelligence, and real-time molecular-level imaging, the procedure aims to ensure the precise reconnection of the spinal cord, nerves, and blood vessels. BrainBridge is actively recruiting top specialists to collaborate on finding solutions to challenges in spinal cord repair and whole-body transplantation techniques. The company emphasizes that its approach is rooted in rigorous scientific research and adheres to the highest ethical standards, committing to engaging in open dialogue with the scientific community, policymakers, and the public to address complex ethical questions and concerns. While the concept of head transplantation may seem like science fiction, it represents a significant step forward in the field of neuroscience, with the potential to transform lives and offer hope to those who need it most.
May 21, 2024, 3 p.m. • WIRED • (6 Minute Read)
Anthropic, an AI startup, has made a breakthrough in understanding the workings of artificial neural networks, which have often been described as black boxes due to the mystery surrounding their inner workings. The research team at Anthropic has managed to peer inside these black boxes, specifically in large language models (LLMs), and identify specific concepts or "features" that artificial neurons signify. By utilizing a technique called dictionary learning, the researchers were able to associate combinations of neurons that evoke specific concepts within the LLM. They have also made progress in manipulating the neural net to augment or diminish certain concepts, with the goal of making LLMs safer and reducing bias. Although the researchers emphasize that they have not completely solved the black box problem and that their techniques may have limitations, their work has shown promising results in shedding light on the inner workings of AI, bringing hope for increased safety and transparency in the field.
May 15, 2024, 7:32 a.m. • The Register • (5 Minute Read)
The news story explores the evolving issue of ownership in software development, particularly when AI is involved in generating code. The article discusses the author's personal experience with using generative AI to expedite coding tasks, highlighting the challenges and ethical questions that arise. The author reflects on the implications of using AI to write code, questioning the ownership of the resulting content and the impact on copyright protection. The article also includes insights from an attorney specializing in Intellectual Property Law, who acknowledges the complexity of determining ownership in code that incorporates AI-generated content. The story emphasizes the need for legal clarity in addressing the potential devaluation of software copyrights due to the integration of AI in coding processes.
May 14, 2024, 7:11 p.m. • Neuroscience News • (5 Minute Read)
Artificial intelligence (AI) computer programs that process MRI results reveal differences in how the brains of men and women are organized at a cellular level, a new study shows. These variations were spotted in white matter, tissue primarily located in the human brain’s innermost layer, which fosters communication between regions. Men and women are known to experience multiple sclerosis, autism spectrum disorder, migraines, and other brain issues at different rates and with varying symptoms. A detailed understanding of how biological sex impacts the brain is therefore viewed as a way to improve diagnostic tools and treatments. The study used an AI technique called machine learning to analyze thousands of MRI brain scans from 471 men and 560 women. The computer programs could accurately distinguish between biological male and female brains by spotting patterns in structure and complexity that were invisible to the human eye. The findings suggest that acknowledging these differences can enhance diagnostic tools and treatments for brain disorders. The results were published online on May 14 in the journal Scientific Reports. The study was funded by the National Institutes of Health and the United States Department of Defense. The researchers involved in the study work at NYU Langone Health and NYU Tandon School of Engineering. They plan to explore the development of sex-related brain structure differences over time to better understand environmental, hormonal, and social factors that could play a role in these changes.
April 19, 2024, 7:15 p.m. • TechTarget • (6 Minute Read)
Artificial intelligence (AI) is a rapidly evolving field with widespread applications in industries such as marketing, customer service, and data analysis. With the constant advancements in AI technology, staying updated on the latest developments and news can be challenging. Fortunately, podcasts offer an effective way to stay current in the AI world. The top 12 AI podcasts recommended by TechTarget cover a wide range of topics, from daily AI news analysis to discussions with expert guests, making them suitable for listeners of all levels. Hosted by long-time tech journalists, industry experts, and researchers, all of these podcasts have received above 4-star ratings and some are award-winning. Whether you're a nontechnical business leader looking to integrate AI into business practices or a tech-savvy professional interested in machine learning and neural networks, there's a podcast tailored to meet your AI needs.
March 21, 2024, 10:35 a.m. • The Register • (4 Minute Read)
In a significant move, Qualcomm has integrated AI support into its Snapdragon 7+ Gen 3 platform, targeting high-end devices at a lower price point than flagship premium smartphones. This new platform offers on-device generative AI capabilities, improved CPU and GPU performance, and Wi-Fi 7 support. Qualcomm's SVP of mobile handsets, Chris Patrick, highlighted that the Snapdragon 7+ Gen 3 promises to deliver new levels of entertainment by integrating next-generation technologies for richer experiences. Notably, the platform includes features such as the Hexagon neural processing unit for generative AI support, a 15% increase in CPU performance, a 45% increase in GPU performance, and support for 5G networks with downlink speeds of up to 4.2 Gbps. This development is expected to impact smartphone makers including OnePlus, Realme, and Sharp in the coming months, as they bring devices to market with the Snapdragon 7+ Gen 3 platform.