OpenAI’s GPT-5 Project Faces Delays and High Costs
OpenAI’s ambitious AI project, GPT-5, codenamed Orion, is experiencing delays and escalating expenses. Despite over 18 months of development, the project has encountered technical challenges, with improvements over existing models like GPT-4 not justifying the substantial financial investment, estimated at around half a billion dollars for a six-month training run. Additionally, data shortages have impeded progress, leading OpenAI to generate new data through human-created content and synthetic means. Internal challenges and competition from rivals have further affected the project’s timeline, leaving the release date of GPT-5 uncertain.
Intensifying Competition for AI Talent
The competition for AI expertise is intensifying, with companies vying for top-tier researchers and infrastructure talent. Databricks recently raised a historic $10 billion, primarily to buy back vested employee stock and provide liquidity. Naveen Rao, VP of AI at Databricks, notes that the talent war extends beyond AI researchers to include individuals skilled in building and managing supporting infrastructure. Major players like OpenAI, Anthropic, Amazon, Google, Meta, xAI, and Microsoft are actively competing for talent, driving up market rates and leading to aggressive recruitment strategies. Rao estimates that fewer than 1,000 researchers globally possess the capability to develop new frontier models, enhancing their leverage and influence in the industry.
AI Integration in Boxing Matches
In a groundbreaking move, the upcoming rematch between Tyson Fury and Oleksandr Usyk will feature an AI-powered judge to unofficially score the bout. This AI judge aims to address criticisms of biased and unfair human judging in boxing by analyzing real-time metrics such as aggression, landed punches, impact, and defensive strategies. While the AI judge’s score will not affect the official result, it provides an alternative evaluation for the public, potentially influencing future boxing scoring methods.
Italy Fines OpenAI for Data Privacy Violations
Italy’s privacy authority, Garante, has fined OpenAI 15 million euros ($15.6 million) for improperly collecting and using personal data through its ChatGPT platform without adequate legal grounding and transparency. The investigation also revealed deficiencies in age verification measures to protect minors from unsuitable content. As part of the penalty, OpenAI is required to conduct a six-month public awareness campaign in Italy about its data practices. This action aligns with global regulatory scrutiny on AI, with entities in the US and Europe examining AI companies and formulating protective regulations, notably the EU’s upcoming AI Act.
AI Cameras to Detect Impaired Drivers in the UK
New AI cameras capable of detecting drunk and drug-impaired drivers are being trialed in the UK. Developed by Australian firm Acusensus, the technology analyzes images of passing cars to identify signs of impairment and alerts nearby police officers. The trial, led by Devon & Cornwall Police as part of their Vision Zero South West campaign, aims to complement police efforts given the extensive road network and limited number of traffic officers. Similar AI cameras have been used to detect motorists using mobile phones or not wearing seatbelts, successfully identifying thousands of violations in various UK regions.
AI Adoption in Australian Corporations
In 2024, numerous Australian companies, including Telstra and major banks, have advanced their use of AI to enhance customer service and operational efficiency. Telstra has deployed AI tools like AskTelstra and One Sentence Summary, improving customer response capabilities and mobile network management. Wesfarmers’ Bunnings implemented “Ask Lionel” to provide real-time product information. National Australia Bank utilizes AI for coding, knowledge management, and marketing, while Macquarie Group introduced Macquarie AI Chat to boost productivity. This trend highlights AI’s transformative potential across different sectors, improving efficiency, customer service, and productivity.
Apple Enhances AI Features in Latest iPhones
Apple has released a free software update, iOS 18.2, for its latest iPhones, adding advanced AI features just in time for the holiday shopping season. This update introduces “Genmojis,” allowing users to create custom emojis quickly, and includes other AI functionalities such as an “Image Playground” for AI-crafted illustrations and email summarization tools. These features are available on iPhones manufactured from 2023 onwards, prompting an expected surge in iPhone upgrades. This move positions Apple in competition with rivals Samsung and Google, while emphasizing user privacy by processing AI data predominantly on the device or in secured data centers.
These developments underscore the rapid evolution and integration of AI technologies across various industries, highlighting both the opportunities and challenges that accompany such advancements.
