ChatGPT Credentials Leak, OpenAI vs Musk

07 March 2024: OpenAI responds to Musk; AI's dark web dilemma, copyright battles, & AI investments, "Machines of Tomorrow" Author Interview

March 07, 2024

teal LED panel

Good Morning. Today's edition unearths a spectrum of AI developments, from Elon Musk's legal battle with OpenAI to unsettling security breaches and pivotal copyright issues. We also shine a light on 'Machines of Tomorrow: From AI Origins to Superintelligence & Posthumanity', where authors dissect AI's trajectory towards superintelligence.

1-Minute Roundup

  • Stolen 225,000 ChatGPT credentials were sold in darkweb.

  • Musk's legal face-off with OpenAI questions the ethos of AI advancement.

  • GPT-4 is under fire for potential copyright breaches.

  • Overjet's $53.2M boost aims to redefine dental diagnostics

  • TollBit's innovative funding model promises a new revenue stream for content creators.

  • Taalas introduces a silicon breakthrough poised to redefine AI efficiency.

  • Rios's $13M infusion to enhance AI robotics

  • Google engineer's alleged tech theft could propel China's AI prowess

  • Accenture's Udacity acquisition aims to create an AI learning powerhouse.

  • Cloudflare's AI "Firewall" sets a new benchmark in securing generative models from cyber threats.

  • Microsoft's upcoming AI-powered Surface devices promise a leap in computing intelligence.

  • AWS's Generative AI Competency program to become a beacon for businesses navigating AI solutions.

  • ChatGPT gains a voice, marking OpenAI's venture into more accessible AI interactions.

Spotlight

ChatGPT Credentials

A staggering 225,000 OpenAI ChatGPT credentials have been compromised and are now circulating in the internet's underbelly, thanks to the sinister efforts of malware like LummaC2, Raccoon, and RedLine. This surge in cyber theft underscores the escalating threat as cybercriminals exploit AI.

  • Malware Menace: Most breaches stem from LummaC2, Raccoon, and RedLine stealer malware, highlighting the critical need for robust cybersecurity defences.

  • Rising Threat: The period from August to September saw a notable spike in compromised devices, indicating a growing interest among cybercriminals in AI-powered platforms.

  • Global Cyber Espionage: Microsoft and OpenAI's revelations about nation-state actors using AI for cyberattacks add a new dimension to the threat, with potential implications for corporate and national security.

How should individuals and organisations safeguard their credentials against such sophisticated threats?

Tug Of War Fun GIF by R Marine Crawley

Elon Musk has initiated a lawsuit against OpenAI, accusing the AI behemoth of straying from its charitable roots to chase profits, primarily through its association with Microsoft. In a dramatic turn of events, OpenAI rebutted, claiming Musk's influence was minimal, and his financial contributions were limited, contrary to his claims of significant involvement and investment.

  • Musk's Move: Alleging that OpenAI, which he co-founded, has prioritised profit over humanity's benefit, Musk's lawsuit challenges the betrayal of open-source ethos.

  • OpenAI's Defense: OpenAI counters Musk's narrative, stating that his $45 million contribution was the extent of his financial involvement. It also highlights internal disagreements over the company's direction, including Musk's proposals to merge OpenAI with Tesla or assume complete control.

  • The Core Contention: The heart of the dispute lies in interpreting OpenAI's mission. Musk insists on a public, open-source commitment, while OpenAI argues for a pragmatic shift towards a for-profit model to sustain its ambitious AGI efforts.

  • Implications for AI: The AI community faces a tension between open innovation and the need for resources to achieve groundbreaking advancements. This legal skirmish sets a precedent for how AI ventures can balance these priorities.

Do you lean towards open-source purity or pragmatic profitability for advancement?

Recent research by Patronus AI highlights leading AI models' knack for unwitting copyright infringement, with OpenAI's GPT-4 topping the charts for regurgitating protected content, pitting innovation against intellectual property rights.

  • Copyright Catch-22: Testing against popular books, GPT-4 was flagged for echoing copyrighted text in 44% of prompts, underscoring the challenges of keeping AI outputs original.

  • Comparative Caution: In contrast, Anthropic's Claude 2 showed more remarkable restraint, reproducing copyrighted content only 16% of the time, suggesting differences in model programming or content filtering.

  • Legal Labyrinth: The findings fuel the ongoing legal tussles between AI developers and content creators, highlighting the complex interplay between AI advancements and copyright laws.

Should AI models have unrestricted access to copyrighted materials for training?

AI Deals

  • Overjet raises a record $53.2M in Series C to enhance its dental AI platform, bringing its total funding to $133M and valuation to $550M.

  • TollBit raised $7M to create a platform that allows AI bots to pay for website content. The company aims to benefit the publishing industry and reduce legal uncertainties.

  • Taalas exits stealth with $50M funding, introducing a silicon AI technology promising a 1000x improvement in AI cost efficiency with its groundbreaking chip.

  • Rios Intelligent Machines secures $13M in Series B to boost its AI robotics in manufacturing, focusing on wood, beverage, and food sectors.

Book Author Interview

In our exclusive interview, discover Pedro Uria-Recio's visionary insights on intertwining humanity and AI. Gain unique perspectives on the future of technology and society from the co-author of 'Machines of Tomorrow: From AI Origins to Superintelligence & Posthumanity.' Read more

Others

  • A Google engineer faces indictment for allegedly filching over 500 files on AI technology, including pivotal info on Google's TPU chips, to aid China-based AI ventures.

  • Amidst escalating AI tech tensions and U.S. chip access curbs, the engineer's alleged espionage could fuel China's AI ambitions, risking a hefty prison term and fines if convicted.

  • Accenture acquires Udacity to spearhead "LearnVantage," a billion-dollar initiative aimed at AI mastery amidst the generative AI surge.

  • Amidst valuation dips, Udacity joins forces with Accenture, shifting from a potential $80M India’s Edtech “Upgrad” deal to bolster global tech skill development.

  • Cloudflare's new "Firewall for AI" targets preempting cyber threats to Large Language Models by vetting incoming prompts.

  • The system integrates with existing web application firewalls, adding layers like rate limitation and sensitive data detection to bolster AI security.

  • Microsoft is gearing up to introduce Surface Pro 10 and Laptop 6 as its pioneering AI PCs on March 21. These PCs boast next-gen processors for enhanced AI performance.

  • The new devices promise a suite of AI features, including the novel AI Explorer for seamless natural language searches across applications.

  • AWS launches Generative AI Competency program to benchmark AI models and tools, aiding businesses in selecting effective generative AI solutions.

  • The program, showcasing over 40 partners at launch, sets a standard for AI expertise and application, promising to streamline the adoption of generative AI technologies for businesses.

  • OpenAI enhances ChatGPT with a Read Aloud function, offering vocal responses in 37 languages across web and mobile platforms, catering to users on the move.

  • The feature, supporting GPT-4 and GPT-3.5, boasts five voice options and language auto-detection, marking OpenAI's foray into multimodal AI capabilities.

Today's stories showcase the complex evolution of AI and its impact on ethical, legal, and societal issues. 'Machines of Tomorrow' offers a contemplative pause, urging us to consider the broader implications of our technological pursuits.