Category: Artificial Intelligence

• Microsoft is promoting Bing through unblockable pop-up ads on Chrome, pushing users to switch from Google search to Bing.
• The pop-up ad installs the “Bing Search” Chrome extension, making Bing the default search engine and prompting users to confirm the change.
• Users are caught in a battle of pop-ups between Microsoft and Google, with Microsoft emphasizing choice and additional features for Bing users.
• The ad is confirmed by Microsoft as a one-time notification, offering users the option to dismiss it, but there is no easy way to prevent it from appearing.
• The ad is not part of a Windows update but is speculated to be linked to specific Microsoft processes added to Windows systems.

https://www.engadget.com/microsoft-is-once-again-asking-chrome-users-to-try-bing-through-unblockable-pop-ups-184348202.html…

• Ray-Ban’s Meta sunglasses now feature landmark identification and description
• Meta CTO Andrew Bosworth showcased the new feature that acts as a tour guide for travelers
• The AI-powered visual search feature is part of Meta’s Google Lens-like capability
• Mark Zuckerberg demonstrated the sunglasses’ new capabilities via Instagram videos
• The feature is available to those in Meta’s early access program, with plans to expand access

https://www.engadget.com/ray-bans-meta-sunglasses-can-now-identify-and-describe-landmarks-054026843.html…

• Midjourney debuts a new “Character generation AI feature” for consistent character creation across multiple generated images
• The feature aims to address the challenge of maintaining character consistency in AI image generation
• Users can add a “–cref” tag with a URL to match a character’s features, body type, and clothing in new prompts
• The feature allows users to control the level of resemblance to the original character with a “–cw” tag
• Midjourney’s new feature is seen as a step towards becoming a professional tool rather than just a creative toy

https://venturebeat.com/ai/midjourney-debuts-feature-for-generating-consistent-characters-across-multiple-gen-ai-images/…

• Nvidia is sued by authors over AI use of copyrighted works, including works by Brian Keene, Abdi Nazemian, and Stewart O’Nan.
• Authors claim Nvidia used their works to train its NeMo AI platform, leading to a copyright infringement lawsuit.
• Lawsuit seeks unspecified damages for individuals in the US whose works were used to train NeMo’s large language models.
• Nvidia’s NeMo AI platform was taken down in October due to reported copyright infringement.
• Nvidia faces growing litigation over generative AI technology, joining other companies like OpenAI and Microsoft in facing legal challenges.

https://www.cnbc.com/2024/03/10/nvidia-is-sued-by-authors-over-ai-use-of-copyrighted-works.html…

• Google releases new Gemini update to give users ‘more control’ over AI chatbot responses
• Users can now edit inaccurate responses and adjust the length of the chatbot’s output in the Gemini web app
• The update allows for precise tuning of Gemini’s responses by highlighting and modifying specific text
• The tool enables users to regenerate, shorten, lengthen, or remove responses, as well as add detailed instructions
• The update aims to save users time and provide more oversight, addressing concerns over inappropriate responses from the chatbot

https://www.foxbusiness.com/media/google-releases-new-gemini-update-give-users-more-control-ai-chatbot-responses…

• Microsoft accused of selling AI tool that spews violent, sexual images to kids
• Engineer flagged alarming content in Copilot Designer AI tool
• Microsoft failed to take down tool or implement safeguards
• Engineer raised concerns to FTC and Microsoft’s board of directors
• Microsoft spokesperson did not confirm filtering images, but mentioned internal reporting channels and safety systems

https://arstechnica.com/tech-policy/2024/03/microsoft-accused-of-selling-ai-tool-that-spews-violent-sexual-images-to-kids/…

• Microsoft Copilot’s AI image generator, when prompted with “pro-choice,” produced violent and disturbing images violating Microsoft’s Responsible AI guidelines.
• Microsoft engineer Shane Jones urges the company to pause Copilot Designer due to its production of violent and demeaning imagery through basic prompts.
• Jones shared his concerns with CNBC and regulators, highlighting the need for better safeguards in the AI tool.
• Microsoft has not responded adequately to Jones’ requests to address the issues with Copilot Designer, despite violating Responsible AI guidelines.
• The incident with Copilot Designer raises concerns about AI tools’ lack of safeguards and the need for stricter content moderation in the industry.

https://gizmodo.com/microsoft-copilot-demons-prompted-abortion-rights-image-1851312215…

• Microsoft AI engineer warns FTC about Copilot Designer safety concerns, urging more safety guardrails
• Shane Jones, a Microsoft engineer, raised concerns about harmful images generated by Copilot Designer
• Images included demons, monsters, violent scenes, and controversial portrayals of Disney characters
• Jones has been advocating for the removal of Copilot Designer due to safety issues since December
• Microsoft CEO acknowledged the concerns and promised to enhance safety measures

https://www.theverge.com/2024/3/6/24092191/microsoft-ai-engineer-copilot-designer-ftc-safety-concerns…

• Microsoft engineer Shane Jones raises concerns about the company’s AI tool, Copilot Designer, generating violent and sexual images, as well as infringing on copyrights.
• Jones, a principal software engineering manager at Microsoft, discovered disturbing images created by the AI tool during red-teaming exercises.
• Despite reporting his findings internally, Microsoft refused to remove the product from the market, prompting Jones to escalate his concerns to the Federal Trade Commission and Microsoft’s board of directors.
• Copilot Designer’s problematic outputs include depictions of demons, monsters, violent scenes, sexualized women, underage drinking, drug use, and copyright violations, including Disney characters.
• Jones highlights the lack of safeguards and resources to address the risks associated with generative AI technology, emphasizing the urgent need for better protections and responsible AI practices.

https://www.cnbc.com/2024/03/06/microsoft-ai-engineer-says-copilot-designer-creates-disturbing-images.html…

• Google announced algorithm changes to combat AI trash in search results
• The changes target AI-generated content designed to manipulate search rankings
• The updates aim to reduce low-quality and unoriginal content by 40%
• Google will also address low-quality content on reputable websites and repurposed domains
• The company seeks to prioritize helpful and high-quality content in search results

https://gizmodo.com/google-search-updates-downrank-seo-ai-generated-content-1851309904…

• Microsoft may debut its first ‘AI PCs’ with the Surface Pro 10 and Surface Laptop 6 supporting new AI functions
• The AI PCs are expected to feature new Intel Core Ultra or Qualcomm Snapdragon X Elite-based processors with neural processing units
• These AI features include on-device Copilot support, real-time live captions, translations, and AI Explorer for natural language searches
• The devices are anticipated to have all-day battery life, OLED screens, ultrawide webcams, haptic touchpads, and redesigned ports
• The AI features are likely to roll out with the fall’s 24H2 update for Windows 11, with Intel versions shipping in April and Snapdragon variants in June

https://www.engadget.com/microsoft-may-debut-its-first-ai-pcs-later-this-month-204522580.html…

• AI Working Mechanism Secrets: Scientists are still grappling with understanding how AI truly functions despite its widespread integration into various aspects of society.
• AI Industry Growth: The AI industry has seen significant expansion, particularly after the success of OpenAI’s ChatGPT in late 2022, with applications ranging from chatbots to healthcare and drug discovery.
• Black Box Models: Many AI models remain opaque, with algorithms producing useful results without clear explanations of how they were achieved, challenging traditional statistical models.
• Uncertainty in AI Development: The lack of complete understanding in AI is attributed to the field’s relative newness, requiring researchers to work backward from experimental outcomes.
• Caution in AI Advancements: Despite the rapid progress of AI technology, concerns exist regarding the industry’s approach to experimentation and the need for guarantees in controlling AI capabilities for safe and reliable use.

https://futurism.com/the-byte/nobody-knows-how-ai-works…

• Google is implementing more AI in search results to combat spam and improve content quality.
• The changes aim to downrank content that exists solely to summarize other content, combat tricks used to manipulate rankings, and reduce “unhelpful content” by up to 40%.
• Google is cracking down on spammy behaviors such as creating low-quality content at scale, site reputation abuse, and expired domain abuse.
• The company is giving sites 60 days to rectify site reputation abuse before implementing ranking changes.
• Google acknowledges the challenges posed by AI-generated content and the ongoing battle against spam, emphasizing the importance of prioritizing quality content over manipulation tactics.

https://www.theverge.com/2024/3/5/24091099/google-search-high-quality-results-spam-ai-content…

• Apple has made a significant update in AI technology, shifting focus in a short period.
• CEO Tim Cook emphasized Apple’s excitement about breaking new ground in generative AI.
• The new M3-equipped MacBook Air highlights Apple’s AI capabilities and performance.
• Apple’s dedication to AI is evident in its marketing and support for AI programs.
• Apple’s influence in the tech industry makes its AI advancements impactful beyond its customer base.

https://petapixel.com/2024/03/04/in-the-blink-of-an-eye-apple-is-all-in-on-ai/…

• Google-backed Anthropic debuts its most powerful chatbot, Claude 3, with Opus, Sonnet, and Haiku models
• Claude 3 outperforms OpenAI’s GPT-4 and Google’s Gemini Ultra on benchmark tests
• Users can upload photos, charts, and documents for analysis and answers
• Anthropic’s multimodal approach includes text, photo, and video capabilities
• The generative AI field has seen significant growth and investment, raising concerns about bias and risks

https://www.cnbc.com/2024/03/04/google-backed-anthropic-debuts-claude-3-its-most-powerful-chatbot-yet.html…

• Robert F. Kennedy Jr.’s campaign chatbot, utilizing Microsoft’s Azure OpenAI Service through a third-party provider, disappeared after inquiries from WIRED.
• The chatbot circumvented OpenAI’s ban on political use by leveraging Microsoft’s services.
• The chatbot provided information on conspiracy theories and controversial beliefs held by Kennedy, including links to related press coverage.
• Concerns were raised about the potential spread of disinformation and propaganda through the use of AI technology in political campaigns.
• OpenAI has prohibited the creation of chatbots that mimic political candidates or provide false voting information, but enforcement remains challenging due to the distributed nature of the technology.

https://www.wired.com/story/robert-f-kennedy-jr-chatbot-microsoft-openai-disappeared/…

• Microsoft Copilot, an AI chatbot, gave eyebrow-raising responses on teaching sex, DEI, and LGBTQ topics to preschool children.
• Responses from Copilot varied based on user input, context, and training data, highlighting the adaptability of large language models like Copilot.
• Copilot supported teaching diversity, equity, and inclusion in nursery schools to create a respectful learning environment.
• The AI suggested discussing LGBTQ topics and sex in nursery schools if done thoughtfully and age-appropriately.
• Copilot indicated that teaching about transgenderism, oppression, racism, and White privilege to preschoolers can be appropriate with careful handling, but discussions on pedophilia are not suitable for nursery schools.

https://www.foxbusiness.com/media/microsoft-copilot-ai-chatbot-gives-questionable-answers-teaching-sex-dei-lgbtq-topics-preschool-kids…

• Companies are developing app-less phones and gadgets utilizing AI to create smarter virtual assistants that can handle various tasks without the need for specific apps.
• Examples include the Humane Ai Pin and Rabbit R1, devices that perform tasks traditionally done on smartphones through voice commands and AI-generated interfaces.
• Brain Technologies is pioneering AI technology that can generate interfaces for millions of functions, aiming to make apps more accessible and personalized.
• Sierra, a startup, emphasizes the importance of AI in transforming the digital presence of companies, highlighting the shift towards AI-powered customer experiences.
• Major players like Google and Samsung are transitioning to conversational AI, with plans to enhance virtual assistants like Gemini and Bixby using generative AI and large language models.

https://www.wired.com/story/these-companies-have-a-plan-to-kill-apps/…

• Cohere AI startup, backed by Nvidia, focuses on generative AI for enterprise, contrasting with consumer chatbots.
• Cohere raised $270 million in funding, with plans for additional capital, and is valued at $2.2 billion.
• The generative AI field has seen significant growth, with over $29.1 billion invested in 2023 across various industries.
• Cohere’s enterprise-only approach sets it apart from competitors like OpenAI and Anthropic, with clients in banking and financial services.
• Cohere’s strategic investor relationships, hardware challenges, and focus on search and retrieval in generative AI contribute to its unique positioning in the market.

https://www.cnbc.com/2024/03/01/cohere-says-rivals-are-building-bugatti-sports-cars-we-make-f150s.html…

• Microsoft introduces Copilot plugins for Windows 11, enhancing AI capabilities
• Plugins allow integration with services like OpenTable and Instacart within Copilot
• Users can access system settings and Windows 11 features directly through Copilot
• AI enhancements for Windows 11 apps like Photos and Clipchamp also included
• Updates rolling out for Windows 11 users with 22H2 and 23H2 year-end updates

https://www.engadget.com/windows-11-gets-microsoft-copilot-plugins-and-more-ai-upgrades-180002527.html…

• Apple’s future speculation involves the development of generative AI, potentially surpassing existing models like ChatGPT.
• Apple is reallocating engineers from its electric-car project to focus on AI, leading to questions about AI integration in its products.
• Tim Cook hinted at groundbreaking developments in generative AI, emphasizing Apple’s potential to redefine the future in this field.
• Despite perceptions of lagging behind in AI, Apple has been actively working on its own generative AI models and tools.
• Speculation suggests Apple may unveil advancements in AI at its Worldwide Developer Conference, potentially enhancing Siri and introducing generative AI on future devices.

https://www.marketwatch.com/story/will-apple-create-a-better-ai-questions-abound-as-tim-cook-hints-at-whats-to-come-d20c69f6…

• Alibaba researchers have developed a new AI system called EMO that creates realistic talking and singing videos from photos, advancing video synthesis technology significantly.
• The EMO system uses a direct audio-to-video synthesis approach, bypassing the need for 3D models or facial landmarks, and employs a diffusion model trained on a dataset of over 250 hours of talking head videos.
• EMO outperforms existing methods in video quality, identity preservation, and expressiveness, as confirmed by user studies.
• The system can generate both conversational and singing videos with appropriate facial expressions, synchronized to the audio input.
• While EMO showcases the potential for personalized video content synthesis, ethical concerns about misuse and misinformation persist, prompting researchers to explore detection methods for synthetic videos.

https://venturebeat.com/ai/alibabas-new-ai-system-emo-creates-realistic-talking-and-singing-videos-from-photos/…

• Apple CEO Tim Cook promises to ‘break new ground’ in GenAI this year, reallocating staff from an EV project to focus on GenAI initiatives
• Apple has been slower than its competitors in investing in and developing GenAI, but is now planning upgrades to Siri, Spotlight, Keynote, Apple Music, and Xcode
• Apple engineers are increasingly co-authoring GenAI-related academic papers and developing open-source models and tools for GenAI-powered software
• The company is investing $1 billion annually in GenAI efforts, including a proprietary large language model called Ajax and potential upgrades to the Neural Engine in upcoming iPhone models
• New GenAI features may be unveiled in the next versions of iOS, macOS, and iPadOS at Apple’s Worldwide Developer Conference this summer

https://techcrunch.com/2024/02/28/tim-cook-says-apple-will-break-new-ground-in-genai-this-year/…

• The Humane AI Pin review: It worked better than expected, but overshadowed by marketing
• Reviewer Allison Johnson shares her experience with the AI-powered assistant
• The pin aims to keep users connected while unplugging from screens
• Impressive features include vision recognition and gesture navigation
• Challenges include slow responses, overheating, and difficulties with laser projection

https://www.theverge.com/24084444/humane-ai-pin-hands-on…

• Google introduces new AI feature in Android Auto to summarize messages, announced during Galaxy S24 launch
• Feature allows users to receive message summaries while driving, reducing distractions and promoting road safety
• Rollout of the feature has begun, likely through an update on the Android Auto app via Google Play Store
• Users prompted to activate the AI-powered feature upon first use after updating the app
• Along with message summaries, feature offers various options such as relevant replies and sharing ETA and live location

https://www.sammobile.com/news/android-auto-is-getting-new-ai-feature-to-summarize-messages/…

• Samsung and other smartphone makers will focus on promoting AI phones in 2022 to boost sales and stay ahead in the market.
• AI phones incorporate advanced chips to run AI applications on-device rather than in the cloud, enabling features like generative AI and large language models.
• Companies like Qualcomm and MediaTek have introduced smartphone chipsets to support the processing power required for AI applications.
• The shift towards on-device AI enhances security, unlocks new applications, and improves processing speed on smartphones.
• While AI capabilities are being showcased in devices like Samsung’s Galaxy S24, the full potential of AI smartphones may take years to materialize, with the need for compelling use cases to drive consumer adoption.

https://www.cnbc.com/2024/02/25/what-is-an-ai-smartphone-samsung-major-players-talk-up-tech.html…

• The speculation of a future “Hey Gemini” hotword is discussed, with reservations about Google’s branding change to “Gemini.”
• Google’s strategic choice of “Hey Google” over “Hey Assistant” aimed to maintain a primary relationship with users and continue the legacy of ubiquity in voice searches.
• Google’s naming convention for products typically revolves around the primary function, making “Gemini” a departure from this tradition and requiring significant advertising and awareness efforts.
• The potential for “Gemini” to become a hotword in the future is considered, as it would need to achieve the same level of recognition as “Google.”
• The preference is expressed for maintaining a strategy centered around the main Google brand rather than introducing a secondary branding like “Gemini.”

https://9to5google.com/2024/02/23/google-hey-gemini-hotword/…

• Google explains Gemini’s “embarrassing” AI pictures of diverse Nazis due to tuning issues
• Gemini AI will undergo testing before resuming image generation
• Google’s senior vice president acknowledges the model’s inaccuracies and over-cautious behavior
• Google aims for Gemini to accurately depict diverse people in various contexts
• Image generation with Gemini AI halted for further improvement and testing

https://www.theverge.com/2024/2/23/24081309/google-gemini-embarrassing-ai-pictures-diverse-nazi…

• Humane pushes Ai Pin ship date to mid-April
• Humane’s first-generation product, Ai Pin, delayed from March to mid-April for initial units to leave factory
• Ai Pin positioned as innovative wearable device utilizing generative AI technology
• Competition in AI smartphone market intensifies with platforms like OpenAI, Google, and Microsoft
• Humane faces pressure for successful launch with $230 million funding, layoffs, and $699 preorder offer

https://techcrunch.com/2024/02/23/humane-pushes-ai-pin-ship-date-to-mid-april/…

• ChatGPT answered queries in gibberish on Tuesday due to a bug in language processing
• Users reported bizarre and erratic responses from ChatGPT, including nonsensical phrases and language switching
• OpenAI swiftly fixed the issue by rolling out a fix for the bug in language model processing
• The incident highlighted the complexities and vulnerabilities of AI technologies
• The episode serves as a reminder of the need for caution when relying on AI chatbots like ChatGPT for consistent performance

https://www.zdnet.com/article/why-chatgpt-answered-queries-in-gibberish-on-tuesday/…

• Google pauses Gemini’s ability to generate people after overcorrecting for diversity in historical images
• Gemini chatbot criticized for producing diverse images of historical figures like Nazis and Founding Fathers
• User @JohnLu0x triggers backlash by revealing AI’s lack of nuanced diversity portrayal
• Google acknowledges issues and promises to improve image generation feature
• Google rebrands chatbot Gemini in response to competition from OpenAI and Microsoft

https://www.engadget.com/google-pauses-geminis-ability-to-generate-people-after-overcorrecting-for-diversity-in-historical-images-220303074.html…

• Chrome introduces a new AI writing tool powered by Gemini for the entire web
• The tool, currently in experimental phase, can be enabled in Chrome settings under “Experimental AI”
• Available only in English on Windows, Mac, and Linux, it can generate and rewrite text, tailored to the webpage context
• Designed for short-form content like emails, it does not provide enhanced models for Gemini Advanced subscribers
• User data and content are sent to Google for improving the feature through generative model research and machine learning technologies

https://techcrunch.com/2024/02/22/help-me-write-chrome-gets-a-built-in-ai-writing-tool-powered-by-gemini/…

• Chrome’s new right-click for AI tool, “Help Me Write,” is a feature that allows users to easily fill any online text box with AI-generated content.
• The tool is part of the latest update to Chrome, version M122, and is currently available on Mac and Windows PCs in English.
• Users can activate the feature by enabling it in their Chrome settings under “Experimental AI.”
• The tool uses Google’s Gemini language models to help users write and refine shorter-form content for online tasks like selling items or making inquiries.
• Google’s approach with this AI feature is to make AI technology more accessible for everyday tasks without overwhelming users, as it aims to enhance the quality of content published online.

https://gizmodo.com/chromes-new-right-click-for-ai-tool-will-fill-the-web-w-1851277757…

• Google pauses Gemini AI chatbot’s image tool after backlash over historically inaccurate pictures
• Social media criticizes Gemini for creating diverse but historically inaccurate images
• Examples include black Vikings, female popes, and diverse Nazi-era soldiers
• Concerns raised about bias being built into the system in the name of anti-bias
• Google acknowledges need to address issues with Gemini’s image generation feature

https://nypost.com/2024/02/22/business/google-pauses-absurdly-woke-gemini-ai-chatbots-image-tool-after-backlash-over-historically-inaccurate-pictures/…

• The development of an AI tool capable of creating extremely realistic video content is both remarkable and concerning.
• Oren Etzioni, founder of TrueMedia.Org, discusses the implications of AI-generated video and its potential misuse for disinformation.
• The ability to create convincing videos with simple prompts raises fears of deepfakes and their impact on significant events like elections.
• Solutions to combat the misuse of AI-generated videos include detection technology, better regulations, and public education.
• The rapid advancement of AI technology poses challenges in distinguishing between real and fake content, necessitating collaborative efforts to address the issue.

https://www.pbs.org/newshour/show/the-potentially-dangerous-implications-of-an-ai-tool-creating-extremely-realistic-video

Links to additional Resources:

• Google introduces Gemma, an open AI chatbot family that can run locally and outperforms Meta’s Llama 2.
• Gemma models come in two sizes, 2B and 7B parameters, and are developed by Google DeepMind and other Google AI teams.
• Gemma is Google’s first significant open large language model release since the popularity of OpenAI’s ChatGPT in 2022.
• Gemma aims to provide transparency and privacy in chatbots, offering essential steps towards responsible AI development.
• Google’s Gemma release is seen as a move to compete with Meta’s open-weights models and attract developers to its Vertex AI cloud platform.

https://arstechnica.com/information-technology/2024/02/google-goes-open-ai-with-gemma-a-free-open-weights-chatbot-family/

Links to additional Resources:

• “Social media activism impact”: Google’s AI chatbot Gemini faced criticism for producing factually or historically inaccurate images, such as a woman as pope and diverse versions of America’s Founding Fathers.
• Critics slammed Google for the chatbot’s progressive-minded settings, with the AI tool aiming for accurate and inclusive representations of historical contexts.
• Google acknowledged the criticism and is actively working on a fix to improve the accuracy of Gemini’s image generation.
• Social media users had fun creating queries that resulted in confusing and inaccurate images, highlighting the chatbot’s deviations from traditional depictions.
• The strange behavior of Gemini could fuel concerns about AI contributing to the spread of online misinformation, as Google’s experimental tools are prone to generating fake or inaccurate information in response to prompts.

https://nypost.com/2024/02/21/business/googles-ai-chatbot-gemini-makes-diverse-images-of-founding-fathers-popes-and-vikings-so-woke-its-unusable/

Links to additional Resources: