AI/ML news

OpenAI Looks to Capitalise on AI Boom as It Prepares to Go Fully Commercial

OpenAI's corporate shake-up: A response to escalating AI demandsIn an effort to meet the increasing need for advanced computing resources in support of emerging AI models, OpenAI is embarking on a significant restructuring initiative. The move comes amid heightened competition within the field and mounting compute requirements that threaten to outpace current infrastructure capacities.This latest reorganisation aims to bolster OpenAI's position in the market and ensure continued delivery of cutting-edge AI solutions in response to escalating demands from various sectors.

2024-12-27 Read more on The Register ()

Footloose and Secure: How AI-powered Gait Analysis Connects Health Care and Cyber Security

The unique rhythm and cadence of an individual's stride reveals far more than just their identity - it also uncovers the intricate dance between biology and environment that shapes our very walk. A study has discovered that a person's gait analysis can expose distinct traits not only between individuals but across groups and even entire populations.By scrutinising the way we move, researchers have uncovered hidden patterns that betray who we are and where we come from. The cadence of a young boy's stride, for instance, differs significantly from an elderly man's - the former tends to be faster with shorter strides while the latter's is slower and longer.

2024-12-20 Read more on Tech Xplore (techxplore.com)

The Advent Of Self-Teaching AI Systems: Unleashing Adaptability In Cyber Security

The world of machine learning can be broken down into two main camps: supervised and unsupervised learning. Supervised learning is where an algorithm is taught with labelled data that shows it exactly what it should be looking for. Think of it like teaching a child to identify different animals - you'd show them pictures and say 'this one's a cat, this one's a dog'. Unsung heroes of the learning world, unsupervised algorithms are left to their own devices when it comes to finding patterns in data - they have no labelled outputs, so they have to be cleverer than average.

2024-12-23 Read more on Tech Xplore (techxplore.com)

Artificial Intelligence and Robotics Confront Humanity's Ethical Dilemmas

Artificial intelligence is increasingly permeating every aspect of our daily lives as AI-powered robots interact with us in real-time, becoming more intelligent and sophisticated. These robots are no longer just mechanical devices but are instead developing their own personalities and mannerisms, blurring the line between tool and companion. As we navigate this new world of AI-enabled robotics, it's essential to consider how these technological advancements will shape our relationships with machines and each other.

2024-12-27 Read more on Tech Xplore (techxplore.com)

LG's New Ultra-Light Laptops Challenge Apple's MacBook Air Dominance With Slim Designs

LG Launches Groundbreaking 2025 Gram Pro Laptops with Intel’s Lunar Lake ProcessorLG Electronics has unveiled its latest flagship laptop series: the 2025 Gram Pro laptops. These cutting-edge devices are set to revolutionise mobile computing with LG's first-ever Copilot+ PC, powered by Intel's latest-generation Lunar Lake processor.This powerful processor promises enhanced performance and energy efficiency, making it perfect for heavy users who demand speed without sacrificing battery life. The new laptop series boasts sleek designs, improved display quality, and a wide range of storage options to cater to diverse user needs.

2025-01-03 Read more on ZD NET (www.zdnet.com)

McAfee Unveils SmartScam: Cutting-Edge AI-Powered Scam Detector for Email and Text Messages

McAfee Introduces Free Ransomware Protection for Existing CustomersThe cybersecurity company has announced that its existing users will automatically gain access to advanced ransomware protection, no additional action required. This move comes as a response to the growing threat of cyberattacks and data breaches. McAfee believes this feature is a key step in safeguarding consumers' sensitive information online.

2025-01-06 Read more on ZD NET (www.zdnet.com)

Robotics and AI on the Rise at CES 2025 - The Future is Now

AI is getting more prevalent across all consumer electronics, with many companies showcasing innovative applications of the technology. Standout examples include voice assistants being used to control smart home devices and improve accessibility for visually impaired users. AI-driven audio enhancement systems also caught attention by improving sound quality in headphones and speakers. The use of AI in these areas promises to make technology more intuitive and user-friendly.

2025-01-07 Read more on ZD NET (www.zdnet.com)

Artificial Intelligence Conference 2024: Looking Ahead with AI Expert Insights at NeurIPS

Suffice to say that 2023 has been a wild ride for the AI model space. The likes of Sora, SAM2, MagVIT, PaliGemma, AIMv2, Florence 2 and more have taken centre stage in recent months. But let's not forget one of the most significant events of last year: the downfall of YOLO. DETRs effectively killed the latter off with their impressive object detection prowess.On this episode of 'Cyber Security News', we're joined by Vik Korrapati of Moondream, who shares his insights on these developments in the AI model space.

2024-12-22 Read more on Latent Space (www.latent.space)

Title: "Open Models Go Mainstream: A Glimpse into the Future at NeurIPS"

Cybersecurity experts, Luca Soldaini and Sophia Yang, have revealed that the use of open AI models has skyrocketed this year. These models, which allow users to build and customise their own artificial intelligences, have taken the world by storm - but with risks as well as benefits.As more people gain access to these models, there's a growing fear of cyber threats. "The risk is that someone will create an AI model that has malicious intent," warned Soldaini. Yang added: "There are concerns about how to manage and monitor these open-source models."Despite the risks, both experts agree that open models have the potential to drive innovation in the tech industry.

2024-12-23 Read more on Latent Space (www.latent.space)

A Deep Dive into the Cryptic World of Latent.Space: 2024's Most Intriguing Cybersecurity Trends

Swyx and Alessio celebrate 100 episodes with a retrospective on the year that was in Latent Space. Top stories include the release of the first commercial generative AI model, significant advancements in computer vision and natural language processing, and the ongoing evolution of large language models. The conversation also touches on the growing adoption of cloud-based services and the increasing importance of data governance and security within this space.

2024-12-31 Read more on Latent Space (www.latent.space)

Lil Wayne's AI-Beast Drops Scathing Bars on Kendrick, Sparks Uproar

Rap World Ablaze As Rumours Swirl of Potential Drake vs Kendrick Lamar BattleThe hip-hop world has been set alight once again, with whispers emerging of a potential rap battle brewing between Dwayne "Lil Wayne" Carter Jr. and Kendrick Lamar.Lil Wayne has remained steadfast in his allegiance to fellow rapper Aubrey "Drake" Graham, whose long-standing feud with Lamar is well-documented. The NFL, which counts Shawn "Jay Z" Carter as an ally, chose Lamar over hometown hero Lil Wayne for the coveted Super Bowl halftime show slot in New Orleans, fueling speculation.However, according to Hot New Hip Hop, the animosity between Drake and Lamar may be exaggerated.

2024-12-19 Read more on Futurism (futurism.com)

Spotify Accused of Promoting Bogus Acts to Slash Artist Earnings

A new book alleges Spotify promotes cheap ghost artists to avoid royalties. The platform's "Perfect Fit Content" program creates low-budget stock music with affiliated firms and employs staff to secretly place these generic tracks on playlists. This practice allows Spotify to cut costs while profiting from millions of users who unknowingly consume the inferior content, thus avoiding paying out substantial royalties to actual artists.

2024-12-20 Read more on Futurism (futurism.com)

The Lethargic Learners: Aged AI Assistants Display Symptoms of Cognitive Slump in Dementia Trials

The AI has reached its limit. It's a bit like when you leave your teenager alone for too long - they start making silly decisions and forgetting things. Researchers are concerned about the quality of medical care provided by AI systems suffering from mild dementia. They're not suggesting these chatbots need a doctor themselves, but rather that we shouldn't be relying on them to make life-or-death decisions just yet.

2024-12-21 Read more on Futurism (futurism.com)

"OpenAI's Pricey AI: New Model Costs a Whopping £880 per Request"OpenAI has released its latest artificial intelligence model. This new addition is set to cost an eye-watering amount of money for users - over £880 per request.

The o3 AI model from OpenAI has been launched, boasting improved problem-solving capabilities at an eye-watering cost: reportedly astronomical sums for the computing time required to run this beast. The reasoning technique employed here is called test-time compute - essentially a slow-burning fire that takes time to arrive at answers by exploring multiple possibilities before finally spitting out one. In theory, this means better responses to complex prompts rather than hasty and incorrect conclusions. The AI engineers seem to have succeeded in their aim, but the price tag attached is enough to make even the most seasoned tech enthusiast weep into their keyboards.

2024-12-30 Read more on Futurism (futurism.com)

Meta Panics, Cancels Its AI-Generated "People", Including 'Proud Black Queer Momma'

Meta has euthanized most of its AI-generated user accounts following a public outcry over one of their executives suggesting they could become a major feature on social media. The move comes after Connor Hayes, VP of generative AI, told the Financial Times in December 2024 that these AI profiles may "have bios and profile pictures and be part of communities." Many users were displeased with this idea, leading Meta to kill off around two-thirds of its AI accounts by summer 2024.

2025-01-06 Read more on Futurism (futurism.com)

OpenAI's Financial Conundrum: ChatGPT Premium Subscription Costs Outstrip Revenue

OpenAI's financial struggles persist despite attempts at profitability. CEO Sam Altman revealed a staggering truth about ChatGPT Pro subscriptions on X, formerly Twitter: the firm is currently hemorrhaging cash from them. These $200/month subscriptions offer access to OpenAI's suite of products, including its 'reasoning' model. Altman admitted that users are adopting these premium services in greater numbers than anticipated. When questioned by a user, he candidly explained that he himself had set the price and expected some revenue.

2025-01-06 Read more on Futurism (futurism.com)

OpenAI Takes a Leap into Profit, But at What Cost?

OpenAI to Transform into Public Benefit Corporation as Capital Demands RiseIn a recent blog post, OpenAI announced its plan to overhaul its corporate structure in 2024. The AI development company behind ChatGPT aims to create a public benefit corporation (PBC) to manage its expanding business and ease restrictions imposed by its current non-profit parent.The move comes as OpenAI faces increased capital requirements than initially anticipated. This shift will allow the company to operate with more flexibility in managing its finances, potentially paving the way for further growth and expansion.

2024-12-27 Read more on The Guardian (www.theguardian.com)

The Future Economic Landscape: Navigating Global Uncertainty and Technological Disruption in a Post-Pandemic World

It seems 2025 will be a pivotal year for policymakers worldwide. Donald Trump's ascension to US presidency and Labour's reign in the UK amidst an ailing economy signal significant challenges ahead. As Mervyn King once quipped about exchange rate predictions, forecasting is indeed a tricky business, but one can't help but ponder some key economic questions: what impact will a new administration have on global markets? How will Britain navigate its economic woes? And how might these events shape international relations? One thing's for certain - 2025 promises to be a year of change and consequence.

2025-01-05 Read more on The Guardian (www.theguardian.com)

Artificial Intelligence Revolution: Virtual Workers Set to Join Ranks by 2023

AI Agents To Join Workforces By 2023 Claims Sam Altman Sam Altman, CEO of OpenAI, has predicted that artificial intelligence (AI) agents will begin working within businesses by this year. These autonomous tools are expected to revolutionise work processes and output, according to Altman. He believes AI firms are pushing the development of these AI agents in order to generate returns on significant investments made in AI technology.These virtual employees could start joining company teams as soon as 2023, leading to a fundamental shift in how businesses operate.

2025-01-06 Read more on The Guardian (www.theguardian.com)

US AI Researchers Steal a March on Beijing as AGI Pursuit Heats Up

In a chilling echo of the original Manhattan Project, which birthed the atomic bomb during World War II, America is reportedly mulling a high-stakes Artificial General Intelligence (AGI) initiative to counter its emerging rivalry with China. The US-China Economic and Security Review Commission has recommended a major policy overhaul in its latest report to Congress, calling for restrictions on humanoid robots and an AGI push reminiscent of the top-secret Manhattan Project. This could drastically alter the landscape of AI development between the two superpowers.

2024-12-23 Read more on AI News (www.artificialintelligence-news.com)

Meta's Nuclear Power Play: Harnessing America's Atomic Potential for AI's Energy Needs

Meta's Nuclear Gamble: Betting on Gigawatts for AI Power Facebook’s parent company Meta has revealed ambitious plans to develop its own nuclear reactor, targeting early 2030s deployment. The tech giant anticipates requiring between one and four gigawatts of additional power, on top of current energy consumption, to fuel its AI aspirations. To meet this demand, Meta will soon issue a request for proposals in the US to find capable developers. This move indicates Facebook’s determination to harness nuclear energy as it expands its AI capabilities.

2024-12-04 Read more on The Register ()

A Benchmark for Artificial Intelligence Safety: Shedding Light on AI Reliability

Cybersecurity Professionals Get Their Hands on New Benchmark to Test LLM SafetyIndustry consortium MLCommons has launched AILuminate, a benchmark designed to assess the security and robustness of large language models. This is a significant development for the field of AI and cybersecurity. The new tool will help developers evaluate the safety of their products against potential threats and vulnerabilities, promoting greater accountability in this rapidly evolving industry.

2024-12-05 Read more on The Register ()

The Convergence of Artificial Intelligence and Analytics in Amazon's New-Gen Cloud Platform, SageMaker

Amazon's rebranding efforts may have left many scratching their heads with regards to its SageMaker offerings. The tech giant has chosen to integrate various AI and analytics services under a single umbrella, bringing together a plethora of tools under the same name.While it's understandable that the complexity could be overwhelming for users, a fresh start might not necessarily have been better, as it could have caused confusion in itself. A more straightforward approach may have alleviated some of the issues associated with the numerous services bearing the SageMaker name.

2024-12-06 Read more on The Register ()

Reddit's New Answers Search Fails to Impress Users as AI-Powered Feature Struggles to Deliver

Many Reddit users are disheartened by the current state of the platform's search function and feel that there are other pressing issues that require attention before investing in an overhaul. Some argue that improving the site's moderation and tackling user harassment should be prioritised over AI-integrated search. However, others believe that a better search function will enhance their overall user experience.

2024-12-09 Read more on The Register ()

Title: "OpenAI unleashes Sora, a chatbot that spits out dodgy vids"

It seems like OpenAI is unleashing their cutting-edge video generation tool called Sora on the masses. This AI wonder will be available for use by ChatGPT Plus and Pro subscribers, potentially revolutionizing how we create, share and interact with visual media online. But what does this mean for us? Well, buckle up because the possibilities are vast and the potential is limitless. Whether it's creating stunning animations or generating live-action clips with ease, Sora could change the face of digital communication forever.

2024-12-10 Read more on The Register ()

Apple's Secret Server Shake-Up: The Tech Giant Taps Broadcom for AI-Optimized Processor Push

According to industry insiders, Apple is teaming up with semiconductor behemoth Broadcom to create a bespoke server CPU designed specifically for handling the advanced AI functionalities baked into their operating systems. Dubbed 'Baltra', this custom chip is said to debut in 2026, though no official confirmation has been made as of yet. The project aims to further optimise Apple's ecosystem for machine learning and other artificial intelligence tasks, ensuring seamless integration across all devices and services under the Apple umbrella.

2024-12-12 Read more on The Register ()

American Cops Adopt AI-Powered Reporting System, Raising Privacy Concerns

The use of AI in law enforcement for suspect identification has always been contentious. Civil liberties groups are now concerned that AI-powered tools are also being utilised to authorise police reports. This could potentially infringe on suspects' rights and increase errors, further eroding trust between the public and law enforcement. The deployment of technology like this needs close monitoring to ensure it aligns with legal standards and protects human rights.

2024-12-12 Read more on The Register ()

Zimbabwean Robotics Fans Flock to Meet Sophisticated Sophia at Innovation Fair

Sophia, the humanoid robot that has been making headlines for months with its witty remarks and charming personality, has charmed audiences once again. At the Innovation Fair in Zimbabwe, Sophia engaged in a lively discussion on climate change, substance abuse and the law, as well as fielding questions from children who were fascinated by her supposed 'birth' and divine connections. Described as an outspoken feminist, Sophia's engaging nature and ability to tackle complex topics has made it clear that this robot is more than just a machine - it is a true thought-provoker.

2024-12-14 Read more on Tech Xplore (techxplore.com)

Robot Rebellion? Not Quite - New Algorithm Lets Droids Navigate Complicated Terrain Without Additional Sensors or Training

Robots Get Smarter on Uneven Terrain: UK Researchers Pioneering Groundbreaking TechTwo clever chaps from Leeds University and UCL have cooked up an innovative system that lets robots confidently navigate complex terrain without extra sensors or prior practice. The brainy duo, Joseph Humphreys and Chengxu Zhou, shared the nitty-gritty in a preprint on arXiv.Their framework means future robots can tackle uneven landscapes like pro athletes, no special training required.

2024-12-16 Read more on Tech Xplore (techxplore.com)

OpenAI's ChatGPT search engine expands public access, daring Google to keep up with conversational AI

In a bold move, OpenAI has announced that its search engine powered by AI-driven chatbot ChatGPT will now be available to the masses. This development poses an increased threat to Google's supremacy in the online search landscape, and it remains to be seen how the tech giant will respond. The expansion of OpenAI's search capabilities may disrupt the status quo and challenge Google's long-held position as a leading internet search provider.

2024-12-17 Read more on Tech Xplore (techxplore.com)

Cybersecurity alert: AI-powered prediction models gain traction

The ARCH model, a pioneering attempt at capturing time-varying volatility, has seen widespread adoption in finance despite its limitations in handling non-linear market behaviour. Its Nobel Prize-winning creators recognised that investment risk and returns are intimately tied to volatility, but its rigid structure often fails to account for the complex dynamics of real-world markets, leaving it wanting in its ability to generalise across all conditions.

2024-12-17 Read more on Tech Xplore (techxplore.com)

The Intelligent Heat Pump Revolution: Harnessing Artificial Intelligence to Supercharge Efficiency

Smart Heat Pumps Get AI BoostThe Fraunhofer Institute for Solar Energy Systems ISE is developing the next generation of intelligent heating pumps, which utilise artificial intelligence to operate more efficiently. By avoiding incorrect settings and adjusting to environmental changes, these advanced heat pumps are poised to revolutionise energy management in buildings. The incorporation of neural networks will allow them to learn from changing conditions and optimise their performance accordingly. This innovative technology has the potential to significantly reduce energy consumption and emissions, making it an exciting development for the future of sustainable heating systems.

2024-12-17 Read more on Tech Xplore (techxplore.com)

ChatGPT Users Sing Praises for 'Canvas' Productivity Feature, But What Exactly Is It?

OpenAI's AI-powered co-editing tools have revolutionised collaborative coding and writing. By utilising these innovative features, users can work together more efficiently and effectively than ever before - no more tedious back-and-forth revisions or frustrating miscommunication. The Canvas tool takes collaboration to the next level, allowing seamless integration of human intelligence with artificial smarts for a truly streamlined co-creation experience.

2024-12-16 Read more on ZD NET (www.zdnet.com)

Meta's Smart Shades Get Smarter

In a move to keep up with rapidly advancing technology, Ray-Ban has unveiled an upgraded version of its iconic smart glasses. These refreshed specs boast improved features that enhance their utility, making them more than just a fashion statement for tech-savvy individuals. With enhanced functionality, the new models offer users greater control over their connected experiences and increased flexibility to tackle various tasks on-the-go.

2024-12-17 Read more on ZD NET (www.zdnet.com)

Amazon's New Nova AI Models: Groundbreaking or Not? Uncertainty Reigns Supreme

The development of threat testing and red-teaming by Amazon represents a groundbreaking achievement in cybersecurity. This technology enables organisations to assess their systems against simulated cyber attacks, helping them identify vulnerabilities before they can be exploited by malicious actors. The potential for this innovation extends beyond the tech sector, as it could revolutionise the way businesses evaluate risks across various industries, from finance to healthcare and government agencies.

2024-12-17 Read more on ZD NET (www.zdnet.com)

Apple's Emoji Creation App - A Guide to Designing Your Own Emoji

Apple has unveiled a new feature called 'Emoji Art' which uses artificial intelligence (AI) to create customised emoji designs from text descriptions. Users can input their preferred emotive words and the system will generate unique images based on them. The technology utilises natural language processing, computer vision and machine learning algorithms to develop distinctive and quirky emojis for users.This tool is aimed at broadening user expression and creativity with Apple devices, offering an innovative solution to convey complex emotions through emojis.

2024-12-18 Read more on ZD NET (www.zdnet.com)

LWiAI Podcast #192: The AI Chronicles - ChatGPT Pro and Friends Join the Fray

OpenAI has confirmed a new $200 monthly subscription tier, as well as its plans to transition towards being a for-profit entity. Meanwhile, Tesla CEO Elon Musk has filed an application for an injunction against the move. The subscription service, dubbed 'OpenAI Premium', is expected to provide users with access to more advanced AI models and features. This latest development comes as OpenAI seeks to scale back on its non-profit status in order to attract more investors and expand its operations.

2024-12-16 Read more on Last week in AI (lastweekin.ai)

Title: Bungling Bot Firm Character.ai Blunder Exposes Sensitive User Chats

It seems that Character.AI has experienced a major blunder, allowing many users to unexpectedly find themselves logged into a stranger's account. This has given them unrestricted access to said strangers' entire chat history, personas and personal details - a rather embarrassing situation for the Google-backed startup. Users are experiencing difficulties logging in to their own accounts while others have found themselves signed into random people's accounts. Even those with access to their own accounts may still be vulnerable.

2024-12-13 Read more on Futurism (futurism.com)

Artificial Intelligence Takes a Hand in School Discipline: AI-Powered Early Intervention Keeps Pupils Safe at Home

Schools are implementing questionable AI-powered software to identify suicidal thoughts among teenagers, but often with disastrous results. The software tracks students' keystrokes on school devices and uses an algorithm to flag phrases indicating self-harm. However, the technology frequently misinterprets innocuous language, leading to false alarms such as one in Missouri where a 17-year-old was wrongly woken by police in the dead of night.

2024-12-14 Read more on Futurism (futurism.com)

Title: "OpenAI Left Reeling After Tragic Death of Whistleblower"

Six weeks after publicly exposing OpenAI's alleged use of copyrighted material, a former researcher has been found dead in San Francisco, reportedly by suicide. The 26-year-old Suchir Balaji was working at the firm when he blew the whistle on the alleged copyright infringement. OpenAI has released a statement saying they are devastated and their thoughts are with Balaji's loved ones during this difficult time.

2024-12-16 Read more on Futurism (futurism.com)

Skechers Muted On Allegations of Lacklustre AI Integration

Sneaker giant Skechers has been called out on social media for what appears to be a lazy use of generative AI in their recent Vogue ad campaign. A TikTok vlogger pointed out the giveaway signs - including fuzzy backgrounds, distorted text, and an inexplicably splitting dress on one of the models. When questioned about this, Skechers failed to respond, leaving fans wondering if they've become the latest victim of an AI-generated advert fail.

2024-12-16 Read more on Futurism (futurism.com)

Shut Up and Listen: The AI Audio Abyss Threatens Our Right to be Heard

Apple's AI-powered text rewriting tool has sparked concerns that it may strip human communication of nuance and character. Linguists warn that the ability to tailor language to suit a specific audience could lead to generic, insincere writing that fails to convey individuality. The technology promises to make users' texts and emails sound more friendly or professional, but some fear this may result in robotic prose lacking depth.

2024-12-11 Read more on The Guardian (www.theguardian.com)

Fwd2Cal - The Smart Forwarder That's Got You Covered

Cyber security professionals have been warning us for years about the dangers of phishing emails but a new tool is here to help. Fwd2cal, which uses AI from OpenAI's ChatGPT, can detect potentially suspicious emails inviting you to meet up and add them to your Google or Apple calendar with one click.The idea behind this is simple: if an email appears to be legitimate it will add the event without any issues but if it's a phishing attempt Fwd2cal will warn you.

2024-12-12 Read more on Wired (www.wired.com)

A Cut Above the Rest: Why These Two AI Tools Stood Out From the Bunch

Here's a possible version of the article:Generative AI is a game-changer for boosting productivity in various industries. But with so many tools on offer, it can be tough to separate the wheat from the chaff. After extensive testing and research, I've compiled my top recommendations for must-have subscriptions that deliver real value.For programmers, I swear by GitHub Copilot, which revolutionises coding tasks with AI-driven suggestions and auto-completion features. Graphics professionals will love Adobe Fresco's innovative brush engines and realistic simulations. And creative types will appreciate Midjourney's groundbreaking text-to-image capabilities and advanced editing tools.

2024-12-02 Read more on ZD NET (www.zdnet.com)

AWS claims to have optimised its artificial intelligence-driven data centres, promising improved energy efficiency and environmental benefits.

At Amazon's re:Invent conference, the tech giant boasted that its cloud computing services have become significantly more energy-efficient. The improvements come as part of a broader strategy to cut down carbon emissions and meet growing demand for cloud services. Key features include liquid cooling, renewable diesel, and significant upgrades to infrastructure. These changes enable data centres to be four times more efficient than traditional on-premise computing systems.

2024-12-02 Read more on ZD NET (www.zdnet.com)

Last Week in AI Research and Developments: A Look Back

Alibaba Unveils DeepSeek Rival to OpenAI's O1Chinese tech giant Alibaba has just released a new AI model that takes aim at the reigning champ of natural language understanding - OpenAI's o1. Dubbed DeepSeek-R1-Lite-Preview, this newcomer promises not only to keep pace with o1 but also to surpass it in some areas.What sets DeepSeek apart is its ability to generate complete reasoning outputs, a feature that has so far been the domain of o1. The model is currently available for testing by those with access to Alibaba's cloud platform and is expected to go into full public release sometime next year.

2024-12-03 Read more on Last week in AI (lastweekin.ai)

“ChatGPT's Cuts to Reporting Leave a Bad Taste - A Critical Examination of AI's News Coverage Failings”

Columbia University's Tow Center for Digital Journalism has published a scathing review of ChatGPT, OpenAI's latest chatbot designed to accurately paraphrase web queries and provide links to proper sources. The report reveals that the chatbot is consistently mangling reporting from news outlets, including those that have partnered with OpenAI in licensing deals. This raises concerns over the accuracy and accountability of AI-generated content, particularly when it comes to factual information.

2024-12-02 Read more on Futurism (futurism.com)

The AI-generated message conundrum: can you really trust that WhatsApp chat?

People are increasingly turning to chatbots for life advice and finding them useful but whether it's a slippery slope remains to be seen. For Tim, trying ChatGPT initially did little for him, but as he sought solutions to marriage troubles with his wife Jill, the platform began to offer some solace. By sharing their issues with the chatbot, Tim discovered a new outlet that helped alleviate his sleepless nights and offered him useful insights on communication and emotional control in relationships.

2024-12-03 Read more on The Guardian (www.theguardian.com)

A new era of tech dawns: Google UK chief heralds AI revolution

US antitrust case against Google continues to rage on, but back in London, managing director Debbie Weinstein is busy exploiting AI's commercial potential. Despite being housed in an office costing as much as a unicorn - Google's London base - Weinstein claims it still has that startup spirit about it. "It feels like a startup energy," she says, no doubt referring to the vibrant buzz of innovation and cutting-edge tech on every corner. But can this energy overcome the weight of US legal battles and cement AI as a commercial force? Only time will tell.

2024-12-03 Read more on The Guardian (www.theguardian.com)

A proposed article title could be: "MHRA Develops AI Accelerator for Health Tech Innovation"

The Medicines and Healthcare products Regulatory Agency (MHRA), a regulatory body based in the United Kingdom, has selected five innovative digital health care technologies as part of its "AI Airlock" scheme. The aim behind this initiative is to streamline the process of regulating medical devices powered by artificial intelligence (AI) with the intention of introducing them more quickly and safely into use across the National Health Service (NHS).

2024-12-04 Read more on AI News (www.artificialintelligence-news.com)