Five Top Tech Takeaways: Apple's AI Gambit, Meta's Forays into AI, Canada Deals with LegalGPT, and Moody's Take on Shell Companies

 

Apple's AI Revolution: iPhone 16 to Feature Cutting-Edge Generative AI

Apple is set to integrate advanced generative AI features in its upcoming iPhone 16 series, leveraging acquisitions and internal developments to enhance AI capabilities, particularly in video compression and large language model efficiency. With significant investments in AI technology, Apple aims to run AI applications directly on iPhone hardware, reducing reliance on cloud services. This move positions Apple as a strong competitor in the generative AI space, with significant reveals expected at the Worldwide Developers Conference, including potential Siri advancements powered by a large language model.

(Source: iPhone in Canada).

Key Takeaways:

• Apple is intensifying its integration of generative AI into the iPhone 16, focusing on in-house AI capabilities and acquisitions.
• The company's goal is to enable AI applications to run directly on iPhones, minimizing cloud dependency.
• Significant advancements, including Siri's potential upgrade with a large language model, are anticipated at Apple's upcoming Worldwide Developers Conference.

Transforming the Future of Connectivity: Meta's Focus on AGI

Mark Zuckerberg, CEO of Meta, is actively entering the race to develop Artificial General Intelligence (AGI), reorganizing Meta's AI research group, FAIR, to align closer with its generative AI product teams. This move aims to leverage Meta's AI breakthroughs directly for its vast user base. Zuckerberg, facing fierce competition for AI talent and resources, emphasizes the importance of generative AI in achieving general intelligence. Meta, boasting significant computing power with a large stock of Nvidia GPUs, is focusing on open-source AI development, contrasting with other companies' more closed approaches. This strategy reflects Zuckerberg's vision of AI's role in future connectivity, blending human and AI interactions across Meta's platforms.

(Source: The Verge).

Key Takeaways:

• Meta restructures to focus on AGI, integrating FAIR with its generative AI product teams.
• Zuckerberg commits to open-source AI development amid intense industry competition for talent and resources.
• Meta's vision includes blending AI with human interaction, enhancing connectivity across its platforms.

Seven Risk Indicators to Identify Shell Companies: Moody's Latest Tool

Moody’s has developed a Shell Company Indicator to aid in detecting financial crimes involving shell companies, identifying seven key indicators of risk: outlier directorships, mass registration, jurisdictional risk, financial anomalies, dormancy, circular ownership, and outlier ages. These indicators help in identifying suspicious behaviors and patterns that may suggest the presence of shell companies, used for illegal activities like money laundering and fraud. The tool is crucial for compliance, risk analysis, and due diligence processes, especially in light of global events like Russia's invasion of Ukraine affecting jurisdictional risk flags. National legislations are also evolving to combat the misuse of shell companies, underlining the importance of tools like Moody’s Shell Company Indicator in the fight against financial crime.

(Source: Moody's).

Key Takeaways:

• Moody's data indicates a staggering 11.5 million outlier directorships, highlighting individuals with an unrealistic number of roles in multiple companies.
• The Shell Company Indicator has identified 4.2 million instances of mass registration and over 655,000 cases of company dormancy, signaling potential shell company activities.
• The tool flags more than 60,000 instances of circular ownership and over 38,000 cases involving outlier ages of beneficial owners, both critical indicators of shell company risk.

AI in the Courtroom: Canada's First Encounter with Fake Legal Cases

A recent incident in a B.C. courtroom marks Canada's first case involving the use of artificial intelligence to create fake legal cases. (This was seen previously in the US; see here.) Lawyers Lorne and Fraser MacLean discovered that opposing lawyer Chong Ke used ChatGPT to prepare legal briefs, unknowingly submitting fictitious cases. This misuse of AI in legal proceedings has raised serious concerns about the integrity of the legal system, highlighting the potential for erroneous judgments and wasted resources. The incident has prompted warnings from legal experts and regulatory bodies, emphasizing the need for lawyers to verify AI-generated content and the potential consequences of misusing such technology in court proceedings.

(Source: Global News).

Key Takeaways:

• AI-generated fake legal cases were discovered in a B.C. courtroom, marking a first in Canada's legal history.
• The incident has sparked concerns about the reliability and misuse of AI tools like ChatGPT in legal research and documentation.
• Legal authorities and experts are warning of the serious implications and potential consequences for lawyers misusing AI technology in legal proceedings.

The Future of Smart Glasses: A Look at Meta AI's Capabilities

The Ray-Ban Meta smart glasses have introduced new AI features, including "multimodal AI" and real-time information updates. While multimodal AI, which responds to queries based on visuals, shows promise, especially in applications like real-time translations and landmark identification, its real-time information accuracy is questionable. Meta AI struggles with current events, often providing incorrect answers. Despite the potential for useful applications, the current version of Meta AI demonstrates significant limitations in reliability and practicality.

(Source: Engadget)

Key Takeaways:

• Ray-Ban Meta smart glasses now feature multimodal AI, allowing interaction based on visual inputs, useful for translations and text summaries.
• The glasses' real-time information capability is currently unreliable, often providing inaccurate responses to basic questions.
• Despite the innovative technology, the practical application and accuracy of Meta AI need significant improvement to be truly useful.

Author: Malik Datardina, CPA, CA, CISA. Malik works at Auvenir as a GRC Strategist that is working to transform the engagement experience for accounting firms and their clients. The opinions expressed here do not necessarily represent UWCISA, UW, Auvenir (or its affiliates), CPA Canada or anyone else. This post was written with the assistance of an AI language model. The model provided suggestions and completions to help me write, but the final content and opinions are my own.

Five Top Tech Takeaways: AI Aces Math Test, NYT sues OpenAI for Copyrights, Samsung's AI Phones, Meta's AI Shift, and EVs Found Frozen in Chicago

DeepMind's AlphaGeometry: A Breakthrough in AI's Math Abilities

Google DeepMind's latest AI system, AlphaGeometry, represents a significant step in AI development by successfully solving complex high-school geometry problems. This advancement, showcased in a Nature publication, indicates a new level of reasoning and planning in AI, a crucial aspect for future artificial general intelligence (AGI). Unlike current generative AI models that struggle with multi-step problems like advanced math, AlphaGeometry was trained on a large, synthetically generated dataset. Though not yet part of Google's Gemini AI model, AlphaGeometry has potential applications in educational settings and has been open-sourced to encourage widespread use and development.

Key Takeaways:

• AlphaGeometry by DeepMind solves high-school level geometry, marking a major AI milestone.
• The system overcomes a common AI challenge of multi-step reasoning and planning.
• Google open-sources AlphaGeometry, paving the way for broader AI educational applications.

(Source: BNN Bloomberg)

New York Times vs. Tech: A Legal Showdown Over AI and Copyright

The New York Times has filed a lawsuit against OpenAI and Microsoft, alleging copyright infringement due to the use of the Times' articles to train their large language models, which power ChatGPT and Copilot. The lawsuit claims that these AI models can produce content that either directly replicates or closely summarizes the Times' articles, impacting the publication's relationship with its readers and financial streams such as subscriptions and advertising. The complaint further asserts that these AI technologies endanger high-quality journalism by undermining news outlets' ability to protect and monetize their content. The Times argues that while the use of its content has been financially beneficial for Microsoft and OpenAI, its attempts to negotiate fair compensation have been unsuccessful. OpenAI has expressed surprise at the lawsuit, noting ongoing discussions with the Times, while Microsoft has not yet responded to the allegations. In addition to seeking damages, the Times is requesting the court to prevent the use of its content in training AI models and to remove its content from existing datasets.

Key Takeaways:

• The New York Times accuses OpenAI and Microsoft of copyright infringement for using its content in training AI models like ChatGPT and Copilot.
• The lawsuit highlights concerns about AI's impact on journalism and the financial implications for news outlets.
• The Times seeks compensation, removal of its content from AI datasets, and a halt to its future use in AI model training.

(Source: The Verge)

Galaxy S24 Series: Samsung's Bid in the AI Smartphone Race

Samsung's Galaxy S24 and S24 Plus have debuted with a focus on AI, incorporating features like search, translation, and message composition enhancements, processed mainly by Samsung's Gauss generative AI model. Despite sporting familiar designs, these models offer improvements in note organization, real-time language translation, and enhanced photo editing, powered by AI. However, some features seem derivative of existing technologies, and there's skepticism about Samsung's commitment to AI as a sustainable innovation rather than a fleeting trend. The hardware updates include a better camera system and slightly larger batteries, while retaining a design reminiscent of previous models.

Key Takeaways:

• Samsung introduces AI-focused features in Galaxy S24 series, emphasizing generative AI capabilities.
• The Galaxy S24's design remains largely unchanged, raising questions about Samsung's innovation focus.
• Skepticism exists over whether Samsung's AI integration is a true advancement or just a trend-following move.

(Source: Engadget)

Tesla Charging Woes in Chicago's Deep Freeze

In Chicago, numerous Tesla vehicles were unable to charge at Supercharger stations during an extreme cold wave, with temperatures dropping to 2F (-19C) and feeling like -20F (-29C) with wind chill. This situation led to several Teslas being towed to local service centers due to their inability to start charging. While cold weather commonly impacts both electric and gas-powered vehicles, this incident highlights a rare case where electric vehicles, specifically Teslas, couldn't charge at all. The issue underscores the challenges that extreme weather can pose for electric vehicle infrastructure and functionality. 

Key Takeaways:

• Extreme cold in Chicago led to Tesla vehicles being unable to charge at Supercharger stations.
• The severe weather resulted in several Teslas needing to be towed for service.
• This incident highlights the impact of extreme temperatures on electric vehicle charging capabilities.

(Source: Electrek)

Meta's AI Pivot: Integrating Teams, Scaling Up GPU Resources

Meta is intensifying its AI initiatives by integrating its AI research and generative AI teams and significantly expanding its GPU infrastructure, with plans to acquire around 600,000 GPUs, including 350,000 from Nvidia, by the end of the year. This move positions Meta among the leaders in technology infrastructure, surpassing Amazon and Oracle's GPU counts. Alongside this expansion, Meta has launched several AI-driven products, such as the Llama language model, AI-enabled ad tools, and a chatbot for Ray-Ban smart glasses. These efforts align with CEO Mark Zuckerberg's focus on enhancing AI capabilities to support Meta's transition towards an AR/VR-centric metaverse.

Key Takeaways:

• Meta is merging its AI research and generative AI teams to bolster its AI product development.
• The company plans to amass a vast GPU arsenal, aiming for around 600,000 units, to support its AI ambitions.
• These developments tie into Meta's strategic shift towards an AR/VR-driven metaverse, as envisioned by Zuckerberg.

(Source: Reuters)

Author: Malik Datardina, CPA, CA, CISA. Malik works at Auvenir as a GRC Strategist that is working to transform the engagement experience for accounting firms and their clients. The opinions expressed here do not necessarily represent UWCISA, UW, Auvenir (or its affiliates), CPA Canada or anyone else. This post was written with the assistance of an AI language model. The model provided suggestions and completions to help me write, but the final content and opinions are my own.

Five Top Tech Takeaways: Bitcoin ETFs, Open AI SOC2 Compliant Offering, GPT Store is Ready, SEC gets Hacked, and AI-Laptops invade CES

The Bitcoin ETF Horse Race is on.

BlackRock’s Bitcoin Fund Makes Its Debut

The U.S. financial market witnessed a significant milestone with the debut of Bitcoin Exchange-Traded Funds (ETFs). Grayscale Bitcoin Trust (GBTC) led the pack with a first-day trading volume of $532 million, closely followed by Invesco Galaxy Bitcoin ETF (BTCO) at $531 million, and Fidelity Wise Origin Bitcoin Trust (FBTC) at $481 million. Other notable ETFs included Hashdex Bitcoin ETF (DEFI) and WisdomTree Bitcoin Fund (BTCW), with BlackRock's iShares Bitcoin Trust (IBIT) placing sixth. This development represents a major shift in the cryptocurrency landscape, offering investors a new avenue to participate in Bitcoin trading through traditional financial instruments.

Key takeaways:

• Grayscale Bitcoin Trust (GBTC) emerged as the leader in first-day trading volume among the newly launched Bitcoin ETFs.
• The launch of these ETFs marks a significant integration of Bitcoin into conventional financial markets.
• BlackRock's iShares Bitcoin Trust (IBIT), despite its industry prominence, ranked sixth in trading volume on its debut day.

(Source: CoinDesk)

OpenAI's GPT Store: A New App Store for the AI Era?

OpenAI has unveiled its GPT Store, a new marketplace designed for the trading of specialized chatbot agents. This platform, which caters to paid ChatGPT users, expands upon the existing ChatGPT Plus service by offering a wider array of tools for purchase and monetization. The GPT Store enables the development of chatbot agents with unique personalities or themes, tailored for specific tasks like salary negotiation, lesson planning, or recipe development. Over 3 million custom versions of ChatGPT have already been crafted, and OpenAI plans to feature notable GPT tools weekly in the store. This initiative is likened to Apple's App Store, encouraging diverse AI development. OpenAI also announced a forthcoming revenue-sharing program, incentivizing creators based on user engagement with their GPTs.

Key takeaways:

• OpenAI's GPT Store provides a platform for buying and selling customized chatbot agents, enhancing the utility of ChatGPT.
• The store offers diverse chatbot applications, from professional tasks to creative endeavors.
• OpenAI is fostering a community of developers with revenue-sharing incentives and legal support for potential copyright issues.

(Source: The Guardian)

OpenAI's Latest Offering: SOC2 Certified ChatGPT Team for Secure Team Collaboration

OpenAI has introduced a new subscription service called "ChatGPT Team," targeting small to medium-sized teams (up to 149 members). This service offers a dedicated workspace and admin tools for managing team interactions with ChatGPT. It includes access to OpenAI's latest AI models: GPT-4, GPT-4 with Vision, and DALL-E 3, along with functionalities to analyze, edit, and extract information from uploaded files. Users can also build and share custom GPTs without coding experience. ChatGPT Team is priced at $30 per user per month or $25 if billed annually, which is more affordable than the enterprise version but costlier than individual subscriptions. The plan promises new features and improvements, and OpenAI assures that team data and conversations will not be used for model training. 

Key Takeaways:

• SOC2 and Non-Leakage of Data: The offering is SOC2 Type I certified, and it will not send data back to OpenAI's LLM. For more on the SOC2 Type I certification, see here. 
• Access to Advanced AI Models: Subscribers get access to the latest OpenAI models, including GPT-4 and DALL-E 3, enhancing their capabilities in text and image generation.
• Cost-Effective for Small Teams: With a pricing strategy aimed at smaller teams, ChatGPT Team is an affordable option for businesses seeking advanced AI tools without the enterprise-level investment.

(Source: TechCrunch)

Bitcoin Chaos Triggered by SEC Account Hack: Understanding the Aftermath

A security breach at the U.S. Securities and Exchange Commission (SEC) caused significant upheaval in the bitcoin market. The SEC's account was compromised, leading to the false announcement of the approval of spot bitcoin ETFs. This unauthorized post caused a 2.5% spike in bitcoin prices, followed by a substantial drop, resulting in a $40 billion value swing. The SEC retracted the statement, confirming the account hack. This incident has raised concerns about market manipulation and regulatory oversight, with the SEC facing pressure from U.S. senators to explain the breach. The Commodity Futures Trading Commission (CFTC) might investigate the matter, given bitcoin's classification as a commodity in the U.S. The SEC and other regulatory bodies are expected to conduct separate inquiries into the hack and any related staff misconduct.

(Source: WIRED)

Key Takeaways

• Impact on Bitcoin Market: The false announcement temporarily influenced bitcoin prices, highlighting the cryptocurrency's sensitivity to regulatory news.
• Regulatory Scrutiny: The incident has prompted U.S. senators to demand explanations from the SEC, stressing the importance of cybersecurity in financial regulation.
• Potential Investigations: The SEC, CFTC, and possibly other regulatory bodies may conduct investigations to understand the breach's circumstances and prevent future incidents.

AI-Powered Laptops: The New Trend at CES 2024

At the CES 2024 consumer electronics trade show in Las Vegas, PC and microchip companies unveiled a new strategy to boost laptop sales: the integration of artificial intelligence (AI). With the pandemic era leading to a saturation in laptop purchases, companies like Advanced Micro Devices (AMD) and Intel are focusing on neural processing units (NPUs) in their latest chip designs. These NPUs are expected to enhance AI functions in laptops, offering a high level of performance with modest power needs. This move is partly aimed at competing with Apple's market share by adding unique AI capabilities to laptops. The inclusion of NPUs is expected to appeal to consumers seeking more advanced, AI-integrated devices.

Key Takeaways

• Introduction of Neural Processing Units (NPUs): The latest chip designs from major companies now include NPUs to enhance AI functionalities in laptops.
• Targeting High-End Laptop Market: This innovation aims to attract consumers towards more expensive, high-end laptops with advanced AI capabilities.
• Competitive Strategy Against Apple: Incorporating AI features into laptops is part of a strategy to compete with Apple and attract consumers looking for cutting-edge technology.

(Source: Reuters)

Author: Malik Datardina, CPA, CA, CISA. Malik works at Auvenir as a GRC Strategist that is working to transform the engagement experience for accounting firms and their clients. The opinions expressed here do not necessarily represent UWCISA, UW, Auvenir (or its affiliates), CPA Canada or anyone else. This post was written with the assistance of an AI language model. The model provided suggestions and completions to help me write, but the final content and opinions are my own.

Five Top Tech Takeaways: AI's 24 Billion Benchmark, OpenAI Revenue Prediction, Google's DuetAI, AI Makes Music and Microsoft's Mobile AI

AI Industry's Explosive Growth: A 24 Billion Visit Phenomenon

The AI industry has experienced remarkable growth, as evidenced by an analysis of the 50 most visited AI tools, which collectively garnered over 24 billion visits from September 2022 to August 2023. ChatGPT dominated this landscape, accounting for 14 billion of these visits, representing 60% of the total traffic. The study, utilizing data from SEMrush, revealed a monthly average growth of 236.3 million visits across these tools, highlighting the rapid expansion and interest in AI technologies. The analysis also showed a significant shift towards mobile usage, with over 63% of users accessing AI tools via mobile devices. However, there was a notable gender disparity among users, with 69.5% being male. The United States led in AI tool usage, contributing 5.5 billion visits, while European countries collectively accounted for 3.9 billion visits. The study underscores the increasing global reliance on AI tools for various applications, from chatbots to image generation.

Key Takeaways:

• The top 50 AI tools saw a staggering 24 billion visits in a year, with ChatGPT leading the pack.
• Mobile devices are the primary means of accessing AI tools, indicating a shift in user behavior.
• The United States is the leading contributor to AI tool traffic, followed by European countries.

(Source: WriterBuddy)

OpenAI's Projected Revenue Growth: Aiming for $5 Billion in 2024

OpenAI has achieved a significant milestone in its financial growth, with its annualized revenue surpassing $1.6 billion. This remarkable increase comes just two months after the company reported reaching $1.3 billion. The surge in revenue began following the launch of ChatGPT Plus, a paid version of its ChatGPT service, in February 2023. Prior to this, OpenAI's primary revenue source was through providing its AI models to developers via an application programming interface. Despite a loss of $540 million in 2022 on a revenue of $28 million, the company's sales have seen a dramatic increase. OpenAI's optimistic projections suggest that its annualized recurring revenue could reach $5 billion by the end of the year, indicating plans for launching new AI services and creating additional revenue streams. The company is also working on a new flagship large language model, GPT-5, which is expected to be more advanced than GPT-4.

Key Takeaways:

• OpenAI's annualized revenue has significantly increased to $1.6 billion, driven by the success of its ChatGPT Plus service.
• The company anticipates its annualized recurring revenue to potentially quadruple to $5 billion by the end of the year.
• OpenAI is developing GPT-5, a new and more advanced large language model, as part of its expansion and innovation in AI services.

(Source: SiliconANGLE)

From Android to iOS: Microsoft's Copilot App Widens Its Horizon

Microsoft has launched its Copilot app for iOS and iPadOS, following its recent introduction on Android. This app provides users with access to Microsoft Copilot (formerly Bing Chat), offering functionalities similar to OpenAI's ChatGPT mobile app. Users can ask questions, draft emails, summarize text, and even create images through an integration with the text-to-image generator DALL-E3. A notable feature of Copilot is that it allows access to GPT-4, the latest large language model from OpenAI, without requiring a subscription. This move signifies Microsoft's shift towards providing a standalone experience akin to ChatGPT, with the company also establishing a separate web experience for Copilot, distinct from Bing.

Key Takeaways:

• Microsoft's Copilot app is now available on iOS and iPadOS, expanding its reach to a wider user base.
• The app offers various AI-driven functionalities, including text summarization, email drafting, and image creation with DALL-E3 integration.
• Copilot provides free access to GPT-4, indicating Microsoft's commitment to offering advanced AI capabilities to its users.

(Source: The Verge)

Navigating the Capabilities of Duet AI in Google Workspace

Google's new AI workplace assistant, Duet AI, is designed to enhance productivity in professional environments. Integrated into Google Workspace, Duet AI offers functionalities such as summarizing documents, drafting reports, and replying to emails. While testing the demo version, it was observed that Duet AI could efficiently handle tasks like summarizing shorter documents and drafting emails for various scenarios. However, the tool demonstrated limitations, including generating inaccurate summaries with fabricated facts and quotes. This "hallucination" issue, common in AI systems, suggests that while Duet AI can be a helpful assistant for routine tasks, it may not be entirely reliable for critical or nuanced work.

Key Takeaways:

• Duet AI efficiently performs tasks like summarizing documents and drafting emails, enhancing workplace productivity.
• The tool occasionally generates inaccurate or fabricated information, a phenomenon known as "hallucination" in AI systems.
• Despite its potential, Duet AI should be used cautiously for important tasks due to its current limitations in accuracy.

Malik's note: Google Bard is still not available in Canada. 

(Source: Business Insider)

 

Microsoft Copilot and Suno: Revolutionizing Music Generation

Microsoft's Copilot AI, an extension of ChatGPT, has recently integrated with Suno, an AI company specializing in music generation. This collaboration allows users, regardless of their musical background, to create personalized songs using simple prompts. The process involves typing commands into Copilot, which then generates complete songs, including lyrics, instrumentals, and singing voices. While this tool is a boon for professionals to experiment with ideas and content creators on platforms like YouTube, Instagram, and TikTok, it's important to note that Suno retains commercial rights to the music produced. Therefore, these AI-generated songs cannot be used for profit-making purposes unless users subscribe to Suno's premium tier, which offers control over commercial rights.

Key Takeaways:

• Microsoft Copilot now enables users to create music effortlessly through its integration with Suno.
• The tool is accessible to anyone, regardless of musical training, and can generate complete songs from simple prompts.
• While the music created is free to use for non-commercial purposes, Suno retains the commercial rights, limiting the ability to profit from these creations without a premium subscription.

(Source: BGR)

Author: Malik Datardina, CPA, CA, CISA. Malik works at Auvenir as a GRC Strategist that is working to transform the engagement experience for accounting firms and their clients. The opinions expressed here do not necessarily represent UWCISA, UW, Auvenir (or its affiliates), CPA Canada or anyone else. This post was written with the assistance of an AI language model. The model provided suggestions and completions to help me write, but the final content and opinions are my own.