OpenAI Introduces Real-Time Search, Kudrow Raises AI Concerns, Broadcom's AI Revenue Surges, UK Coalition Demands Stronger AI Copyrights, and more...
12:23 Share

Good morning. It's December 17th and this is your daily brief in AI. Here's everything you need to know.

OpenAI is making waves with the global rollout of its new search feature for ChatGPT, set to launch on December 17, 2024. This new functionality allows real-time web searches and delivers direct conversational answers, potentially reducing reliance on traditional search engines. This move positions OpenAI as a formidable competitor in the search engine market, which is still dominated by Google.

The announcement was made during OpenAI's 12 Days ShipMass event, where they also unveiled a premium subscription tier called ChatGPT Pro, priced at $200 per month. The new search feature boasts enhanced categorization for specific queries, providing context and explanations, along with the capability to generate various types of creative content, making it a versatile tool for users.

Additionally, the search functionality improves the presentation of information across categories such as weather, news, sports, maps, and stocks, offering users a more organized experience. However, AI search engines are not without challenges. Issues such as misinformation risks and copyright concerns have led to litigation from traditional news media against OpenAI.

A recent study highlighted concerns regarding the accuracy of chat GPT search, revealing significant problems with source attribution and a failure to provide accurate citations in many tested cases. The reliance on training data raises further concerns about outdated or incorrect information, urging users to verify important facts. As OpenAI continues to innovate, the competitive landscape is evolving.

Tech giants like Google are responding by launching their own AI-driven solutions, including the Gemini chatbot. Google's announcement of Gemini 2.0 aims to develop intelligent agents capable of addressing user inquiries, reflecting the increasing pressure on traditional search models.

Experts suggest that OpenAI's integration of live web data with conversational AI may push Google to innovate beyond its advertisement-dependent business model. This rollout comes amid scrutiny of OpenAI, particularly following controversies related to copyright issues and the recent death of a former researcher. Concerns are mounting in Hollywood about the implications of artificial intelligence in the film industry, particularly with Tom Hanks' upcoming movie, Here.

Lisa Kudrow has voiced her worries that advancements in AI could lead to job losses for future actors. Hanks has also commented on the potential for AI to recreate performances indefinitely, raising significant ethical questions about authenticity in entertainment. Directed by Robert Zemeckis, HERE employs digital de-aging technology, enabling Hanks and Robin Wright to portray their characters over multiple decades.

The film, inspired by Richard Maguire's graphic novel, chronicles the lives of Richard and Margaret from their youth into their 80s. With a budget of $45 million financed by Miramax, Here showcases advancements in aging technology that contrast sharply with older films requiring extensive visual effects.

The film is currently competing with 20 others for a spot on the Oscars shortlist for Best Visual Effects, highlighting the industry's increasing interest in AI-driven tools. This development has sparked discussions about the need for AI protections, especially following the actors' and writers' strikes in 2023, which were partly driven by concerns over AI's impact on jobs.

Industry leaders are expressing a mixture of optimism and caution regarding AI's role in Hollywood, anticipating both cost savings and potential job losses. Hanks described the digital de-aging process as "kooky," allowing for immediate playback of younger versions of actors during filming, which adds a unique dynamic to the production. Zemeckis defends the use of AI, stating it is meant to enhance storytelling rather than replace human performers.

Additionally, Lionsgate's collaboration with AI Startup Runway to develop generative AI models signifies a broader trend towards AI integration in film production. Broadcom has reported a staggering 51.2% year-over-year revenue increase in its fiscal Q4 2024 earnings, reaching $14.1 billion.

This remarkable growth was significantly bolstered by a 150% surge in AI revenue, which totaled $3.7 billion.

Following this announcement, Broadcom's shares rose over 6% on December 11, particularly after revealing a collaboration with Apple to develop an AI server chip, expected to enter mass production by 2026. The company's stock has seen a remarkable 101.4% increase in 2024, indicating robust demand in the semiconductor market.

In contrast, NXP Semiconductors, which focuses on automotive chip solutions, has encountered difficulties. Its stock has declined by 5.6% in 2024 due to cyclical downturns in key markets. NXP's third-quarter revenue fell 5.2% year-over-year to $3.25 billion, primarily attributed to inventory destocking in the automotive sector, which makes up 56% of its sales.

Looking ahead, NXP anticipates a continued revenue decline in Q4 2024, forecasting between $3 billion and $3.2 billion, with adjusted earnings per share expected to fall below analyst estimates. Despite these challenges, NXP has committed to returning value to shareholders with a planned dividend distribution of $1.0114 per share.

Broadcom also has a strong track record in this area, having distributed $22 billion in fiscal 2024 and recently increasing its dividend by 11% to 59 cents per share.

The broader semiconductor landscape remains optimistic, with Bank of America predicting a 15% growth in industry sales, reaching $725 billion in 2025, although this is a slight decline from the 20% growth expected in 2024.

Deutsche Bank forecasts that semiconductor revenue growth will accelerate to approximately 16% year-over-year in 2025, with AI-related companies expected to lead this growth. Memory sales are projected to surge by 20% in 2025, following a significant 79% increase year-over-year in 2024. A coalition advocating for copyright and generative AI policy has outlined three key principles aimed at shaping the future of content creation.

These principles include establishing a dynamic licensing market with strong copyright protections, ensuring control and transparency for content creators, and fostering growth and innovation across the creative and tech sectors. The importance of creators' rights in AI development cannot be overstated, particularly as the creative industries contribute over £100 billion annually to the UK economy.

In light of this, stakeholders are urging the UK government to develop a robust copyright framework that not only preserves creators' rights but also supports long-term growth in both the creative and tech sectors. At a recent committee meeting, Culture Secretary Lisa Nandy addressed questions regarding the government's potential adoption of an opt-in or opt-out model for protecting creatives involved in AI development.

Nandi reassured attendees that the government is committed to ensuring proper remuneration for creators and is focused on finding the right balance in upcoming consultations. This initiative aligns with findings from a report by CISAC, which warns that generative AI could negatively impact songwriters and composers by up to £10 billion over the next decade. Nandi indicated that the consultation process is imminent, although the government has yet to finalize its approach to AI and copyright issues.

Recent polling conducted by Reset Tech and YouGov shows strong public support for transparency in AI model training, with 72% of respondents believing that AI companies should pay royalties to content creators whose works are utilized for training. The coalition proposes a dynamic licensing system that includes strong accountability measures for generative AI developers to ensure compliance with UK copyright law.

Members of this coalition include notable organizations such as BPI, PRS for Music, PPL, MPA, and UK Music, representing a wide array of interests. They envision a collaborative future where creative industries and generative AI developers work together, promoting mutual growth in the online marketplace while emphasizing the necessity for a robust copyright framework to safeguard creators' rights.

The year 2024 has proven to be a transformative period for artificial intelligence, marked by significant advancements across various domains. Breakthroughs in large language models, computer vision, and AI safety are at the forefront of this evolution, underscoring the rapid progress in the field. Among the standout developments are 10 influential research papers that have sparked innovation and practical applications, making them essential reading for both researchers and AI enthusiasts.

These papers reflect the cutting-edge trends in AI, encouraging exploration and experimentation in their methodologies. One notable innovation is QEN2, developed by Alibaba, which boasts a modular AI architecture optimized for multimodal tasks, excelling in generating text, images, and code.

Another significant contribution comes from the GEMMA models, which focus on embedding safety and fairness into AI systems through novel training techniques aimed at reducing bias and enhancing robustness. The ORCA-LLM enhances reasoning capabilities via example-based training, making it suitable for applications such as AI tutors and data analytics tools. Google's Gemini 1.5 is also noteworthy, capable of managing long context processing in natural language.

Processing with a context length of 10 million tokens. Ideal for document analysis. ChatGPT++ further advances in context learning. Enhancing personalization and coherence in user interactions, which is particularly beneficial for customer support and educational platforms.

In computer vision, Vision Mamba introduces state-space models that achieve competitive performance with linear complexity, making them suitable for low-latency applications. The MixR-A-7B model employs a mixture of expert architecture for dynamic resource allocation, improving efficiency in multitasking scenarios.

Additionally, CLAW-LM enhances coherence in long-form text generation through context aggregation techniques, proving useful for academic research summaries and customer support. Lastly, Kernel-Arnold networks combine kernel methods with deep learning to improve interpretability and robustness in complex data tasks, with applications in time series analysis and scientific research.

These developments collectively illustrate the rapid advancements and the potential future directions in AI technology. This has been your daily brief in AI. To read more about these stories, follow the links in the episode bio. You can also subscribe to these updates via email at www.brief.news. For more daily podcasts about the topics you love, visit brief.news forward slash podcasts. Tune in tomorrow. We'll be back with everything you need to know.