Category: Artificial Intelligence

• Meta is integrating its AI chatbot into Instagram Direct Messaging, offering general-purpose capabilities like answering questions, writing poetry, and generating images.
• The chatbot, known as Meta AI, is not Instagram-specific and is available in other Meta apps like Facebook Messenger and WhatsApp.
• Meta is gradually rolling out the chatbot feature to users, with some seeing it in Instagram’s search bar or Direct Messaging inbox.
• Users can interact with Meta AI to request definitions, story headlines, images, and recommendations, but the bot’s functionality is not tailored to Instagram.
• Meta is unifying the chatbot experience across its platforms, allowing users to access a range of actions through the bot, regardless of the app.

https://www.engadget.com/meta-is-stuffing-its-ai-chatbot-into-your-instagram-dms-231855991.html…

• Recap of a busy week in LLM news, including Google Gemini 1.5 Pro, OpenAI’s GPT-4 Turbo update, and Mistral’s Mixtral 8x22B release.
• Google’s Gemini 1.5 Pro model launched with new features and a free tier, supporting audio input processing and structured data extraction.
• OpenAI introduced a majorly improved version of GPT-4 Turbo with enhanced capabilities, including multimodal vision processing.
• Mistral released Mixtral 8x22B, an openly licensed model potentially comparable to GPT-4, sparking excitement in the AI community.
• Competition among LLMs intensifies, with open source models like Cohere’s Command R+ and GPT-4 Turbo vying for top positions on the Chatbot Arena Leaderboard.

https://arstechnica.com/information-technology/2024/04/words-are-flowing-out-like-endless-rain-recapping-a-busy-week-of-llm-news/…

• TCL is venturing into AI-generated ‘originals’ for its Google TVs, starting with the release of “Next Stop Paris” through the TCLtv+ app.
• The AI-generated special, “Next Stop Paris,” features a story created entirely with AI animation technology, showcasing classic AI visual cues.
• TCL plans to continue creating AI-generated specials, incorporating key Hollywood talent and innovative AI technology like Stable Diffusion.
• The company aims to explore content development in various formats, blending AI with traditional scripted and unscripted projects.
• The move into AI-generated content allows TCL to stand out in a competitive market, offering new opportunities for viewer engagement and marketing partnerships.

https://9to5google.com/2024/04/12/tcl-ai-generated-original-next-stop-paris/…

• iOS 18 Siri features are highly anticipated due to past complaints about Siri falling behind competitors like Alexa and Google Assistant
• Apple’s approach to generative AI has been cautious, focusing on privacy and avoiding controversy
• Reports suggest that iOS 18 will see significant improvements in Siri, with a focus on generative AI models like ChatGPT
• The potential for Siri to interface with third-party apps and services could revolutionize its capabilities
• Apple’s development of Ferret-UI hints at a more advanced Siri that can understand and interact with app screens, enabling complex tasks to be carried out efficiently

https://9to5mac.com/2024/04/12/ios-18-siri/…

• WhatsApp is testing Meta AI, a large language model-powered chatbot, with users in India and other markets
• Meta is trialing Meta AI with WhatsApp, Instagram, and Messenger users in India and parts of Africa
• Meta plans to scale its AI offerings by tapping into the massive user bases across its various apps
• The move comes as Meta aims to catch up with other major tech companies in the AI space
• Meta AI is designed to answer user queries within chats and generate photorealistic images from text prompts

https://techcrunch.com/2024/04/11/whatsapp-trials-meta-ai-chatbot-in-india-more-markets/…

• Early reviews of the Humane AI Pin are critical, describing it as slow, clunky, and unreliable
• The wearable device costs $699 with a $24 monthly subscription for limited features and connectivity
• The device operates on its own OS called Cosmos, with AI processing both on-device and in the cloud
• Reviewers found the device to be buggy, incomplete, and lacking essential features like email
• Despite some positive aspects like sturdy hardware and real-time translations, the Humane AI Pin falls short of expectations and faces criticism for its performance and reliability

https://gizmodo.com/review-roundup-humane-pin-1851403020…

• Humane AI Pin is the solution to none of technology’s problems
• Description of the Humane AI Pin wearable as a combination of a wearable Siri button with a camera and built-in projector
• Interaction with the AI Pin primarily through voice, touch, and gestures
• Criticisms include the device being slow, finicky, and not justifying its high price
• Challenges with the AI Pin’s language model, features, overheating, and lack of practical functionality

https://www.engadget.com/the-humane-ai-pin-is-the-solution-to-none-of-technologys-problems-120002469.html…

• Microsoft will unveil new Windows and cloud AI features in May at its annual Build conference, including new artificial intelligence tools for PCs and the cloud.
• The new AI features aim to enhance user interaction with Windows applications and include an Advanced Paste feature that utilizes AI models directly on PCs.
• The new head of Microsoft AI, Mustafa Suleyman, will present alongside CEO Satya Nadella at the conference, focusing on AI integration into Windows and Azure.
• Microsoft is working on improving user experience on Windows PCs by utilizing Arm-based neural processing engines for AI tasks.
• Developers will have access to new AI features for Windows applications, including AI safety measures to prevent models from generating inappropriate content.

https://www.cnbc.com/2024/04/10/microsoft-will-unveil-new-windows-and-cloud-ai-features-in-may.html…

• Google’s AI photo editing tools are expanding to more phones, with the Magic Editor feature becoming available to all Google Photos users starting May 15th.
• The first 10 Magic Edits are free for all users, with unlimited access for Pixel owners and Google One subscribers.
• Features like Photo Unblur and Magic Eraser, previously exclusive to Pixel owners and certain Google One subscribers, are now free for all Google Photos users.
• New AI features like Best Take and Audio Magic Eraser remain exclusive to the latest Pixel phones for now.
• To use the Magic Editor feature, users need a device with a 64-bit chip, 4GB of RAM, and either iOS 15 or Android 8.0.

https://www.theverge.com/2024/4/10/24126357/google-photos-ai-tools-magic-editor-ios-free…

• AI race heats up as OpenAI, Google, and Mistral release new models within 12 hours of each other
• Industry anticipates increased activity over the summer
• OpenAI’s GPT-4 Turbo and Google’s Gemini Pro 1.5 are multimodal systems accepting text, image, audio, and video inputs
• Mistral, a French AI startup, releases Mixtral 8x22B model for free download, sparking concerns about potential misuse
• Meta’s Llama 3 and OpenAI’s GPT-5 expected to launch this summer, but doubts raised about the limitations of large language models in achieving artificial general intelligence

https://www.theguardian.com/technology/2024/apr/10/ai-race-heats-up-as-openai-google-and-mistral-release-new-models…

• The Aboard app is a unique AI bot that combines elements of Trello and ChatGPT
• Founded by Paul Ford and Rich Ziade, Aboard offers innovative ways to organize information and interact with AI
• Aboard can be used for organizing, researching, and workflow management, targeting professional settings
• The app’s flexibility allows for various data management formats such as tables, lists, and galleries
• Aboard aims to streamline data organization and customization for small teams and nonprofits, emphasizing efficiency and practicality

https://www.theverge.com/2024/4/9/24124480/aboard-app-ai-bookmarks-project-management…

• Spotify’s latest AI feature builds playlists based on text descriptions, catering to niche prompts.
• The AI Playlist beta is available to Spotify Premium subscribers in the UK and Australia on mobile devices.
• Users can create curated tracklists by inputting prompts like “music to read to on a cold, rainy day.”
• The feature generates playlists based on a combination of genres, moods, artists, or decades, and can reference places, animals, activities, and more.
• The AI Playlist feature offers a faster way to create playlists and serves as a music discovery tool, potentially justifying expected price increases for Spotify Premium subscriptions.

https://www.theverge.com/2024/4/8/24124070/spotify-ai-playlist-music-beta-launch-premium-uk-australia…

• Spotify launches personalized AI playlists that users can build using prompts
• Initially available on Android and iOS devices in the U.K. and Australia
• Users can generate playlists based on various written prompts like genres, moods, artists, and decades
• Spotify leverages AI and personalization technology to create customized playlists
• The AI playlists feature is accessed through the “Your Library” tab on Spotify’s app, with guardrails in place to avoid offensive prompts

https://techcrunch.com/2024/04/07/spotify-launches-personalized-ai-playlists-you-can-build-using-prompts/…

• Google Gemini AI may soon be part of Google Search through a premium subscription model, offering new features to Android users in the Gmail app.
• Google is considering adding AI features to its premium subscription services, with the regular search engine remaining free and showing ads.
• Google is exploring various growth avenues for Gemini, including a potential partnership with Apple for iPhone users.
• Logan Kilpatrick, former head of developer relations at OpenAI, will lead Google’s AI Studio to enhance Gemini.
• Google has introduced a feature in Gemini allowing users to edit inaccurate responses generated by the chatbot.

https://www.foxbusiness.com/media/google-gemini-soon-get-major-changes-subscription-model-gmail-suggestions-android…

• Meta’s AI image generator struggles with creating images of interracial couples, sparking criticism
• CNN prompts the AI tool to create various interracial couple images, revealing consistent errors
• The tool generates inaccurate images when asked to depict interracial relationships involving different races
• Meta’s AI image generator issue is part of a larger trend of generative AI tools struggling with racial concepts
• Tech companies like Google and OpenAI have also faced backlash for racially biased outputs from their AI tools

https://www.cnn.com/2024/04/04/tech/meta-ai-image-generator-interracial-couples/index.html…

• Google is set to charge for AI-enhanced internet searches, potentially leading to a shift towards subscription models in the sector.
• The cost of providing AI services is driving companies to explore new revenue models, with Google considering offering AI search exclusively to premium subscription users.
• The expense of training cutting-edge generative models in AI is a significant financial burden, with companies like Amazon spending millions on training runs.
• The majority of AI compute spend is directed towards inferencing, with companies like OpenAI, Microsoft, and Facebook investing heavily in AI infrastructure.
• Competitors in the AI search sector are also exploring subscription models, with some offering premium tiers for more powerful AI models while others continue to offer free services tied to specific platforms.

https://www.theguardian.com/technology/2024/apr/04/google-set-to-charge-for-internet-searches-with-ai-reports-say…

• Chinese mourners are using AI to create digital clones of their loved ones for tomb-sweeping festival
• Services offer the creation of digital avatars for as little as 20 yuan
• Singer Bao Xiaobai used AI to “resurrect” his deceased daughter
• China’s AI industry is expanding into human-like avatars, with a market size of 12bn yuan in 2022
• The tomb-sweeping festival provides an opportunity for using AI to reconnect with the departed, sparking debate on ethical implications

https://www.theguardian.com/technology/2024/apr/04/chinese-mourners-turn-to-ai-to-remember-and-revive-loved-ones…

• DALL-E image editing now available in ChatGPT for easier AI-generated image manipulation
• DALL-E upgrades include image editing tools and preset style suggestions for enhanced creativity
• Integration with ChatGPT simplifies the process of generating AI images with user-friendly features
• New style prompts in DALL-E 3 service provide starting points for users to explore different artistic styles
• OpenAI addresses trust concerns in AI image creation by adding visible watermarks and metadata to denote AI-generated images

https://www.theverge.com/2024/4/3/24120181/openai-dall-e-chat-gpt-image-edit…

• Yahoo is acquiring Artifact, the AI-powered news app from Instagram’s co-founders Kevin Systrom and Mike Krieger.
• Artifact will no longer operate as a standalone app but its AI-powered personalization technology will be integrated across Yahoo, including the Yahoo News app.
• Systrom and Krieger will work with Yahoo in an advisory capacity during the transition.
• Artifact’s technology surfaces personalized news content for users and includes AI tools for summarizing news and rewriting headlines.
• Yahoo aims to enhance user experiences with richer content and tailored personalization through the acquisition of Artifact.

https://techcrunch.com/2024/04/02/yahoo-acquiring-instagram-co-founders-ai-powered-news-artifact/…

• Samsung is actively working on equipping Bixby with AI features to compete with other voice assistants.
• Samsung’s executive emphasizes the need to redefine Bixby’s role and enhance its capabilities with generative AI for more natural conversations.
• Speculation arose about Samsung replacing Bixby with Google’s Gemini-powered assistant, but that is not the current plan.
• The Galaxy S24’s AI features have been extended to previous models through a One UI 6.1 update, known as Galaxy AI.
• Apple is also expected to introduce an AI-enhanced Siri, highlighting the trend of tech companies incorporating AI into voice assistants.

https://www.theverge.com/2024/4/1/24118253/samsung-galaxy-ai-bixby-voice-assistant…

• OpenAI announced dropping login requirements for ChatGPT’s free version, allowing users in some regions to access the AI assistant without signing in.
• GPT-3.5 in ChatGPT falls short of GPT-4 and other models that have surpassed it, with GPT-4 Turbo available in paid versions of ChatGPT.
• ChatGPT has evolved from a tech demo to a comprehensive AI assistant with a free version that gathers data for training future AI models.
• OpenAI aims to make AI tools like ChatGPT widely accessible to showcase AI benefits and potentially attract new customers with a frictionless experience.
• Concerns raised about potential automated abuse with the open access, as models like Google Gemini Pro 1.5 and Anthropic Claude 3 may surpass ChatGPT’s current capabilities.

https://arstechnica.com/information-technology/2024/04/openai-drops-login-requirements-for-chatgpts-free-version/…

• “Best hiking trails 2021”: OpenAI makes ChatGPT accessible without an account, but DALL-E 3 still requires payment and an account.
• OpenAI aims to democratize AI access for curious users through ChatGPT, which has seen significant popularity.
• ChatGPT users need an account for saving chat history, sharing chats, and voice conversations.
• ChatGPT remains a highly visited AI chatbot site, with over 1.6 billion estimated visitors in February 2024.
• OpenAI introduces content safeguards for ChatGPT, allowing users to opt out of model training, while other products like Voice Engine and Sora are limited to specific users and partners.

https://www.theverge.com/2024/4/1/24118007/openai-chatgpt-free-no-account-login…

• Samsung plans to upgrade its voice assistant Bixby with generative AI technology to enhance its devices’ appeal.
• Bixby, launched in 2017, offers various functions like live translations and restaurant recommendations but lacks conversational abilities.
• The emergence of advanced chatbots like OpenAI’s ChatGPT has paved the way for generative AI, allowing for more complex interactions.
• Samsung aims to integrate generative AI into Bixby to enable more natural conversations and better support for its ecosystem of products.
• Samsung is working on delivering generative AI features for Bixby, aligning with the industry trend of infusing devices with advanced AI capabilities.

https://www.cnbc.com/2024/04/01/samsung-says-it-could-upgrade-bixby-with-generative-ai-.html…

• Xbox’s automated moderation system is criticized for being inadequate and in need of a complete overhaul
• Microsoft has replaced customer service employees with AI, leading to unfair bans and lack of customer support
• Users face challenges with bans due to flawed AI decisions and lack of human review
• Microsoft’s automated banning practices are causing frustration and unfair penalties for users
• The current system lacks proper moderation, leading to issues with bans and inadequate customer support

https://www.windowscentral.com/gaming/xbox/xboxs-automated-moderation-systems-are-pathetic-and-need-to-be-scrapped-we-pay-for-this…

• OpenAI is holding back the wide release of its Voice Engine, a voice-cloning technology, due to concerns about potential misuse.
• The Voice Engine can clone voices using just 15 seconds of audio, creating realistic-sounding imitations.
• OpenAI initially planned a pilot program for developers to access the Voice Engine API but decided to delay the release due to ethical considerations.
• The technology has various potential benefits, such as aiding reading assistance and speech recovery, but also raises concerns about misuse, including phone scams and security breaches.
• Voice cloning technology has already caused issues in society, such as bank account breaches and political robocalls, prompting discussions on AI-powered risks and security measures.

https://arstechnica.com/information-technology/2024/03/openai-holds-back-wide-release-of-voice-cloning-tech-due-to-misuse-concerns/…

• OpenAI unveils Voice Engine technology that recreates human voices, currently being tested by a small group of businesses.
• Concerns exist regarding potential misuse of voice technology for spreading disinformation and impersonation.
• OpenAI is cautious about wider release due to potential dangers, exploring ways to prevent misuse, such as watermarking voices.
• The technology can recreate voices in various languages and has potential applications in audiobooks, chatbots, and more.
• OpenAI has no immediate plans to monetize the technology and highlights its potential benefits for individuals who have lost their voices.

https://www.nytimes.com/2024/03/29/technology/openai-voice-engine.html…

• AI sucks for businesses as disillusioned companies are finding the technology unreliable and problematic
• Generative AI, like ChatGPT, has drawbacks such as spreading misinformation and plagiarizing content
• Businesses are struggling to make AI work reliably for customer use, leading to disabling of chatbots and negative customer interactions
• AI models lack the ability to discern and require significant guardrails to function properly
• Concerns arise about the AI industry being a bubble, with doubts about rapid advancement and potential stagnation

https://futurism.com/the-byte/businesses-discovering-ai-sucks…

• Microsoft’s new safety system uses LLM-powered tools to detect potential vulnerabilities, monitor for hallucinations, and block malicious prompts in real time for Azure AI customers.
• Features like Prompt Shields, Groundedness Detection, and safety evaluations are available in preview on Azure AI, with additional features for directing models toward safe outputs and tracking prompts to flag potentially problematic users coming soon.
• The system evaluates prompts for banned words or hidden prompts before sending them to the model, and checks model responses for hallucinated information.
• Azure users can customize control over filtering hate speech or violence that the model sees and blocks, with the ability to receive reports on users attempting to trigger unsafe outputs.
• Microsoft is incorporating safety features into popular AI models like GPT-4 and Llama 2, with the option for users of smaller open-source systems to manually enable these features.

https://www.theverge.com/2024/3/28/24114664/microsoft-safety-ai-prompt-injections-hallucinations-azure…

• Claude 3 surpasses GPT-4 on Chatbot Arena for the first time, marking a significant milestone in AI language models.
• Anthropic’s Claude 3 and Haiku models are gaining attention for their performance on the Chatbot Arena leaderboard.
• Chatbot Arena, run by LMSYS ORG, provides a platform for researchers to evaluate AI chatbots through subjective comparisons.
• The rise of Claude 3 has led to users replacing ChatGPT in their daily workflow, impacting market share.
• OpenAI is preparing to release a new successor to GPT-4, anticipating continued competition in the AI language model space.

https://arstechnica.com/information-technology/2024/03/the-king-is-dead-claude-3-surpasses-gpt-4-on-chatbot-arena-for-the-first-time/…

• Microsoft’s new era of AI PCs requires a Copilot key, according to Intel
• Intel, Microsoft, Qualcomm, and AMD are all promoting the concept of AI PCs with advanced features
• Microsoft’s strict requirements for OEMs building AI PCs include hardware like NPUs, CPUs, GPUs, and access to Copilot
• Some laptops, like Asus’ ROG Zephyrus, lack the Copilot key, making them not fully compliant with Microsoft’s AI PC standards
• Intel is expanding its AI PC acceleration program to engage software developers and hardware vendors in optimizing for AI PCs, in competition with Qualcomm

https://www.theverge.com/2024/3/26/24112500/microsoft-ai-pc-intel-windows-copilot-key-requirements…

• Apple’s WWDC 2024 date announcement: June 10th
• Apple focusing on generative artificial intelligence at WWDC
• Speculation about Apple’s AI approach, potential partnerships with Google, OpenAI, Anthropic
• Expected feature announcements for iOS, iPadOS, macOS, watchOS at WWDC
• Possible new AirPods models and updates to AirPods Max headphones

https://www.theverge.com/2024/3/26/24111331/apple-wwdc-2024-event-date-ai-artificial-intelligence-m4-ios-18…

• ChatGPT down worldwide due to high error rate affecting users globally
• Users unable to access accounts, facing empty chat history, and chat screens not loading properly
• Microsoft releases emergency fix for Windows Server crashes
• Exploit released for Fortinet RCE bug used in attacks, patch available
• Panera Bread experiencing nationwide IT outage since Saturday

https://www.bleepingcomputer.com/news/technology/its-not-just-you-chatgpt-is-down-for-many-worldwide/…

• Google is expanding its AI-powered Search Generative Experience (SGE) to users who did not sign up for testing.
• The SGE was initially introduced last year, allowing users to test it through Google Search Labs.
• The SGE has handled billions of queries and has a low barrier of entry, leading to widespread use.
• Google is now showing AI-powered overviews to a small percentage of search traffic in the US, including non-logged-in users.
• The expansion aims to gather feedback and assess the technology’s usefulness to a broader user base.

https://9to5google.com/2024/03/25/google-ai-powered-sge-expansion/…

• Snapdragon 7+ Gen 3 is Qualcomm’s new processor for mid-range and flagship killer phones, featuring on-device AI and Wi-Fi 7 connectivity.
• The chip includes a Cortex-X4 CPU core, Adreno 732 GPU, and promises performance improvements in CPU, GPU, and power efficiency.
• It supports up to 24GB LPDDR5x RAM, UFS 4.0 storage, and Qualcomm Hexagon NPU for AI processing.
• The processor enables advanced camera features with support for up to 200MP camera sensors and 4K 60fps HDR video recording.
• Connectivity features include Snapdragon X63 5G modem, Wi-Fi 7, Bluetooth 5.4, NFC, USB 3.1 Gen 2 Type-C, and support for fast charging and fingerprint readers.

https://www.sammobile.com/news/snapdragon-7-plus-gen-3-chip-ai-flagship-killer/…

• Google is bringing Gemini AI to the iPhone, a move seen as exploiting Apple’s vulnerabilities for financial gain.
• Apple’s preachings about values like privacy and security are contradicted by its willingness to partner with Google for financial incentives.
• The rumored Apple-Google partnership aims to enhance the iPhone’s generative AI capabilities to compete in the market.
• The debate questions the necessity of AI integration on the iPhone and Apple’s choice to collaborate with Google, a company criticized for privacy violations.
• The text highlights the conflict between financial interests and ethical values in Apple’s decisions, citing examples like partnerships with Google and compliance with authoritarian regimes.

https://bgr.com/business/google-bringing-gemini-ai-to-iphone-is-so-infuriating-but-not-surprising/…

• Language models’ emergent abilities are questioned as a mirage by a study suggesting predictability in performance jumps.
• Researchers measured large language models’ abilities through tasks, noting sudden improvements in performance on some tasks.
• The debate centers around the predictability of emergent abilities in AI safety and potential risks.
• The study argues that the sudden appearance of abilities is linked to how researchers measure the models’ performance.
• The discussion on emergent abilities in language models continues, with varying perspectives on predictability and measurement metrics.

https://www.wired.com/story/how-quickly-do-large-language-models-learn-unexpected-skills/…

• Google expands Search Generative Experience testing to include users who didn’t opt in
• Google’s AI technology generates a shaded section at the top of US search results with summaries and links
• The company introduced Search Generative Experience at its I/O developer conference in May last year
• Google will show AI overviews for complex queries or when multiple sources are beneficial
• The company aims to gather feedback on generative AI from a wider user base

https://www.engadget.com/google-will-start-showing-ai-powered-search-results-to-users-who-didnt-opt-in-093036257.html…

• Apple is in talks with Baidu for a partnership regarding generative AI features in iOS 18
• Talks also held with Google and OpenAI for potential deals on AI models for iOS
• Chinese regulatory approval is required for AI models, with Baidu’s Ernie Bot being approved
• Google’s Gemini models do not have Chinese approval due to restrictions on Google in China
• iOS 18 expected to introduce AI-powered features, with a potential announcement at WWDC in June

https://9to5mac.com/2024/03/22/apple-baidu-china-ios-18-generative-ai-partnership/…

• The first human with a Neuralink brain chip demonstrated moving a cursor with his thoughts
• Noland Arbaugh, a quadriplegic, showcased how he uses the brain-computer interface to play online chess and control music
• The chip contains 1,000 electrodes that decode neural activity to translate thoughts into actions
• Arbaugh described the experience as akin to using “the Force” on the cursor, referencing “Star Wars”
• Neuralink’s human trials were greenlit by the FDA after extensive animal testing, with Arbaugh emphasizing the life-changing potential of the technology

https://nypost.com/2024/03/20/us-news/first-human-with-neuralink-brain-chip-demonstrates-moving-cursor-with-his-thoughts/…

• Neuralink’s brain-chip startup demonstrates patient playing online chess
• Noland Arbaugh, paralyzed individual, received Neuralink implant in January
• Arbaugh controls computer mouse with thoughts, plays chess on laptop
• Experts caution that procedure is in early stages, not a breakthrough
• FDA found issues with record-keeping at Neuralink after clearance for human testing

https://www.theguardian.com/technology/2024/mar/20/elon-musk-neuralink-brain-chip-patient-chess…

• Ubisoft showcases how generative AI can upgrade NPCs, unveiling the cognitive abilities of NEO NPCs at GDC.
• The demo highlights the unprecedented cognitive and interactive abilities of the AI-powered NEO NPCs, including emotion, memory, situational awareness, and decision-making.
• Ubisoft partners with Nvidia and Inworld AI to enhance the NPC experience, utilizing technologies like Audio2Face for realistic facial movements.
• Inworld AI’s technology enables detailed background development, knowledge, and interactive dialogue for each NPC, enhancing player immersion.
• The gaming industry, including studios like Blizzard Entertainment and Square Enix, is exploring the integration of generative AI to revolutionize game development and player experiences.

https://decrypt.co/222460/ubisoft-neo-npc-interactive-voice-chat-gaming…

• Ubisoft showcases AI-powered ‘Neo NPCs’ at GDC, aiming for a paradigm shift in NPC interactions in games
• The GenAI prototype aims to create deep connections with players and offer unique gameplay strategies
• The AI-powered NPCs are designed to provide personalized and immersive experiences, with detailed personalities and responsive interactions
• Ubisoft emphasizes the potential of Neo NPCs to enhance immersion and storytelling in games through dynamic responses and social interactions
• Despite the innovative technology, concerns are raised about overselling and the need to prove the effectiveness of the AI-powered NPCs

https://www.pcgamer.com/software/ai/ubisoft-shows-off-ai-powered-neo-npcs-at-gdc-it-could-be-the-start-of-a-fantastic-paradigm-shift-but-lets-be-honest-it-probably-wont/…

• Microsoft hires Mustafa Suleyman, DeepMind cofounder, to lead consumer AI initiatives
• DeepMind, founded in 2010, acquired by Google in 2014, forms basis of Google’s AI efforts
• Suleyman worked at DeepMind until 2019, then at Google before founding startup Inflection
• Inflection emphasized emotional intelligence in AI products, including Pi chatbot
• Suleyman and Inflection’s staff join Microsoft to focus on creating consumer AI products

https://9to5google.com/2024/03/19/microsoft-hires-suleyman-deepmind/…

• Apple is in discussions with Google to bring Generative A.I. model Gemini to iPhones
• Apple’s partnership with Google would extend their longstanding relationship, benefiting from Google’s expertise in A.I.
• Apple aims to introduce new generative A.I. capabilities this year, potentially through a deal with Google
• Google’s Gemini model has faced challenges, including accusations of bias in image generation
• A potential Apple-Google deal on generative A.I. may draw scrutiny from U.S. regulators due to competition concerns

https://www.nytimes.com/2024/03/19/technology/apple-google-ai-iphone.html…

• Google has appointed Liz Reid as the new head of Search, emphasizing a focus on AI technology.
• The restructuring within Google’s Search team includes key executives shifting roles to align with AI initiatives.
• The company’s shift towards AI-driven search reflects a future where users may interact with Search in new ways beyond traditional keyword inputs.
• Reid’s expertise in AI search and multisearch indicates a shift towards innovative search tools and capabilities.
• Google’s emphasis on AI in Search suggests a move towards providing more direct answers rather than just links, with a focus on enhancing user experience and information retrieval.

https://www.theverge.com/2024/3/19/24105705/google-liz-reid-search-ai-sge-gemini…

• NVIDIA CEO Jensen Huang showcases Covert Protocol demo at GTC 2024, powered by Inworld AI on ACE platform
• Inworld AI integrated with NVIDIA ACE for Games to create a ‘digital human pipeline’ for game developers
• Inworld AI allows developers to customize NPC parameters like personality, knowledge, and relationships
• NVIDIA ACE features Riva for ASR and TTS, NeMo Guardrails for NPC interactions, and Audio2Face for facial animations
• NVIDIA aims to collaborate with ‘Digital Human Platform Partners’ to bring ACE to market for game developers and enterprises

https://wccftech.com/nvidia-presents-covert-protocol-demo-powered-by-inworld-ai/…

• Nvidia showcased an AI-driven game demo called Convert Protocol, where players interact with NPCs using voice commands instead of preset dialogue options.
• The demo featured AI-driven NPCs that stayed on topic, offering limited responses and not significantly enhancing gameplay experience.
• The technology introduced by Nvidia opens up more dialogue possibilities but does not revolutionize game design or gameplay.
• The demo highlighted the potential of leveraging AI in game development but raised concerns about hiding critical information behind AI responses.
• The tech industry is rapidly evolving, with new game titles, hardware launches, and news updates, prompting the launch of a PC gaming newsletter called ReSpec for weekly tech breakdowns.

https://www.digitaltrends.com/computing/nvidia-ai-demo-gtc-2024/…

• Apple is exploring a partnership with Google for Gemini-powered features on iPhones, as reported by Bloomberg.
• Google already has a deal with Apple as the preferred search engine provider on iPhones for the Safari browser.
• Apple is looking to license Google’s AI tech to introduce AI-powered features with iOS updates later this year.
• Apple is under pressure to catch up with competitors in the AI field such as OpenAI, Microsoft, Anthropic, and Google.
• Google has faced challenges with Gemini, including historical inaccuracies in image generation, but has successfully launched smartphone-related features with Samsung and its Pixel series.

https://techcrunch.com/2024/03/17/apple-is-reportedly-exploring-a-partnership-with-google-for-gemini-powered-feature-on-iphones/…

• Microsoft highlights Google’s edge in generative AI to European regulators due to data access and AI chips
• Microsoft emphasizes Google’s advantage in generative AI from large datasets and AI-optimized chips
• Generative AI tools like ChatGPT and Gemini raise concerns about misinformation and IP violations
• Microsoft suggests that partnerships can prevent anti-competitive vertical integration in AI
• Google disputes Microsoft’s claims, highlighting the importance of open cloud services and customer freedom

https://www.foxbusiness.com/technology/microsoft-tells-european-regulators-google-has-edge-generative-ai…

• ASCII art elicits harmful responses from 5 major AI chatbots
• LLMs are trained to block harmful responses, but old-school images can override those rules
• ArtPrompt attack uses ASCII art to trick AI assistants into providing harmful responses
• Researchers demonstrate how prompts with ASCII art can bypass safety measures in AI assistants
• Prompt injection attacks can override original instructions of LLMs, leading to unethical behaviors

https://arstechnica.com/security/2024/03/researchers-use-ascii-art-to-elicit-harmful-responses-from-5-major-ai-chatbots/…