YouTuber Sues OpenAI for AI Data Scraping
?? Exciting things are happening! We’ve got all the latest updates, including amazing AI news and important industry moves. We’re thrilled to bring you the best of what’s new in tech!
Get cozy and join us as we dive into the world of data and technology! ???
OpenAI, Anthropic, and Meta: Tracking the lawsuits filed against the major AI companies????
Disruptive technologies often face legal challenges, and the early stages of artificial intelligence have been particularly litigious.
As companies like OpenAI, Alphabet, and Meta compete to lead AI development, there's been a surge in lawsuits. Many content providers accuse these companies of using their intellectual property without permission to train large language models. Some lawsuits focus on privacy concerns, while others, like those from Elon Musk, claim that AI companies have strayed from their original missions.
Copyright lawsuits
Currently, over 25 copyright lawsuits are pending against AI companies. One notable case, Thomson Reuters v. ROSS Intelligence, involves allegations that ROSS used data from Thomson Reuters' Westlaw platform to train its AI, with ROSS claiming fair use. The trial, initially set for August 23, has been postponed, but it could set a legal precedent for similar cases.
Other significant lawsuits include:
Meta is going to let preteens use Horizon Worlds????
Meta is expanding access to Horizon Worlds, its 3D social platform, to users aged 10 to 12, as announced on Wednesday. Parents will have substantial control over the experiences their preteens can access, with experiences receiving age ratings to guide parents in making appropriate choices for their children.
Preteens will need a parent-managed Meta account to use Horizon Worlds and must request access to the platform. Once approved, they can either request access to specific experiences or parents can enable a setting that allows them to explore all experiences within a certain age rating. Meta will categorize experiences as suitable for ages 10+, 13+, or 18+.
Meta has implemented additional protections for preteens. Voice chat is disabled by default on parent-managed Meta accounts, but parents can enable it for specific contacts. Preteens will appear offline by default unless a parent changes this setting. Additionally, Horizon Worlds' personal boundary feature will be activated for parent-managed accounts, preventing other users from getting too close in the virtual space.
Horizon Worlds was previously only available in VR, but now some experiences can also be accessed via the Meta Horizon mobile app, which Meta has recently redesigned to emphasize Horizon Worlds.
Last month, Meta announced that preteens with parent-managed accounts can, with parental approval, chat or call with others in mixed reality and virtual reality experiences.
Threads deepens its ties to the open social web, aka the ‘fediverse’????
Threads is strengthening its connections with the fediverse, also known as the open social web, which includes platforms like Mastodon, Pixelfed, PeerTube, and Flipboard. On Wednesday, Meta announced that Threads users can now view fediverse replies on posts beyond their own. Additionally, posts made through the Threads API, including those from third-party apps and scheduling tools, will be syndicated to the fediverse.
This update was first announced via an in-app message stating that API posts would be shared starting August 28.
Since the Threads API launched in June, integrations with platforms like Hootsuite, Sprinklr, Sprout Social, and Grabyo have made Threads more valuable for brands and social media marketers. This expansion is crucial for extending the reach of high-profile accounts, such as @potus.
In contrast, Elon Musk’s X has restricted API access over the past year by eliminating its free tier and increasing prices to boost revenue for the platform formerly known as Twitter.
Today's major update to Threads involves how fediverse replies are displayed. Since June, users could see replies to their posts if they enabled fediverse sharing. Now, users will also see fediverse replies on others' posts, bringing more content into Threads without needing a large increase in its user base.
OpenAI Sued by YouTuber for AI Data Scraping????
Millette's lawyers, in a lawsuit filed on Friday in the U.S. District Court for the Northern District of California, allege that OpenAI secretly transcribed Millette's and other creators' videos to train their AI models for ChatGPT and other generative tools.
The complaint alleges that OpenAI significantly benefited from the creators' work by collecting their data, while also violating copyright law and YouTube's terms of service, which prohibit using videos for non-platform-related applications.
The complaint alleges that OpenAI enhanced its products by incorporating plaintiffs' videos into its language models, benefiting subscribers who paid for these improvements. However, the plaintiffs and potential class members did not receive any compensation.
The lawsuit argues that it is unfair and unjust for OpenAI to retain the benefits derived from the plaintiffs' work without restitution.
The proposed class action seeks restitution for unjust enrichment and alleges unfair competition under California law.
In addition to class certification, the complaint demands damages, monetary relief, injunction relief, and reasonable attorney fees and costs. OpenAI has not yet responded to requests for comment.
Training AI
AI models that generate content, like those from OpenAI, do not possess true intelligence. Instead, they use pattern recognition to predict data occurrences based on context, analyzing a wide range of content such as videos and essays.
Most models are trained on data from public websites and online datasets. Businesses argue that fair use allows them to collect and use this data for commercial model development.
However, many copyright holders disagree and are pursuing legal action to halt this practice. With other data sources declining, video transcriptions have become essential for training these models.
OpenAI's Legal Troubles
Lawsuits challenging OpenAI's use of internet data for ChatGPT training continue to pose a legal challenge for the company.
Publishers are increasingly accusing these companies of misusing confidential data and seeking compensation for their work. Meta and OpenAI have argued to the US Copyright Office that distributing copyrighted content online is protected under fair use laws, as it is considered "publicly available."
Despite this, they will need to defend their position in court as they face multiple lawsuits related to copyrighted material.
Ransomware Gang Leaks Data Allegedly Stolen From Microchip Technology????
The Play ransomware group has leaked gigabytes of data allegedly stolen from US semiconductor supplier Microchip Technology. Microchip had previously disclosed a cyberattack disrupting operations at some of its manufacturing facilities, impacting order fulfillment. The group listed Microchip on their Tor-based site on August 27 and began leaking over 5GB of data on August 29, including personal information, IDs, and various business documents.
The hackers assert that the leaked data is only a portion of what they've stolen and threaten to release more if Microchip does not pay.
The Play ransomware group, active since June 2022, has previously targeted organizations like A10 Networks and Rackspace.
Last year, government agencies in the US and Australia issued warnings after the cybercriminals targeted hundreds of organizations. The group’s website now lists over 550 victims.
South Korean hackers exploited WPS Office zero-day to deploy malware????
The South Korean cyberespionage group APT-C-60 has exploited a zero-day vulnerability in the Windows version of WPS Office to deploy the SpyGlace backdoor on targets in East Asia.
Developed by the Chinese company Kingsoft, WPS Office is a popular productivity suite with over 500 million active users worldwide. The zero-day flaw, identified as CVE-2024-7262, has been used in attacks since at least late February 2024 and affects versions from 12.2.0.13110 (August 2023) to 12.1.0.16412 (March 2024).
Kingsoft quietly fixed the vulnerability in March without notifying customers of its active exploitation, prompting ESET—who uncovered the campaign and the flaw—to release a detailed report today.
领英推荐
Alongside CVE-2024-7262, ESET's investigation also revealed a second critical flaw, CVE-2024-7263, which Kingsoft addressed in late May 2024 with version 12.2.0.17119.
APT-C-60 exploitation
CVE-2024-7262 is a vulnerability in how WPS Office handles custom protocol handlers, like 'ksoqing://,' which can execute external applications via specially crafted URLs in documents. The flaw arises from inadequate URL validation, allowing attackers to create malicious hyperlinks that lead to arbitrary code execution.
APT-C-60 exploited this by embedding malicious links in MHTML spreadsheet files disguised under decoy images. When clicked, these links execute a base64-encoded command that loads a malicious DLL (ksojscore.dll) designed to download a backdoor named 'SpyGlace' from the attacker's server.
Attackers Exploit Critical Atlassian Confluence Flaw for Cryptojacking ????
Threat actors are still exploiting a critical remote code execution (RCE) vulnerability in Atlassian Confluence, first discovered in January, to turn targeted cloud environments into cryptomining networks.
Trend Micro has identified two distinct cryptojacking attacks using this flaw, tracked as CVE-2023-22527, which affects Confluence Data Center and Confluence Server. This server is used for enterprise-level deployments of Atlassian's collaboration and documentation platform.
Rated 10 out of 10 on the Common Vulnerability Scoring System (CVSS) when discovered, the bug has proven highly exploitable, with cryptojacking now added to its list of potential attacks. Despite Atlassian's patching of the flaw eight months ago, attacks persist, as detailed in a Trend Micro blog post from August 28.
The attacks involve techniques such as deploying shell scripts and XMRig miners, targeting SSH endpoints, eliminating competing cryptomining processes, and using cron jobs to maintain persistence, according to Trend Micro's senior engineer, Abdelrahman Esmail.
Trend Micro has also found thousands of other exploitation attempts related to CVE-2023-22527 in recent months and urges users to patch their servers immediately if they haven’t already done so.
Microsoft, Google in AI Proxy WarRead ????
In today’s tech-driven world, where the creators are often overshadowed by those with deep pockets, just two words—artificial intelligence (AI)—can make a company highly appealing. Major tech giants have invested an astonishing $12 billion (with more promised) into just two firms: OpenAI and Anthropic. Recently, Google poured $2 billion into Anthropic, while Amazon pledged an additional $4 billion to address the funding disparity that saw Microsoft invest $10 billion in OpenAI.
This investment in OpenAI made headlines last November when ChatGPT was launched, showcasing a chatbot that could reshape creative processes. However, the Wall Street Journal reports that Google's deal with Anthropic will initially fulfill only a quarter of its commitment, with terms still undisclosed. Despite this, it's clear that companies leading the AI charge are reaping significant rewards.
AI isn’t new and others can use LLM too
This doesn't mean there aren't other players in the AI field doing noteworthy work, though some may overstate its novelty. Recently, a former colleague pointed out that a self-switching turn indicator on a car is also a form of artificial intelligence. If AI is so advanced, why isn't this technology available for motorcycles?
Since both OpenAI and Anthropic focus on large language models, it's likely that new startups will continue to introduce incremental improvements. The frenzy among Big Tech companies to invest in AI suggests a strong desire to either be involved in or own existing AI-driven platforms. Being left out of these innovations could jeopardize their future relevance.
What are these AI companies gaining??
So, do these AI startups, including OpenAI (which might not fit neatly into that category), all address the same problems? Not exactly. Anthropic, for instance, is focused on AI models that may not seem as immediately intriguing to the general public but are highly relevant to enterprises looking for sustainable growth. In contrast, OpenAI is rapidly developing new features, such as voice and image generation with its AI models.
Anthropic, on the other hand, emphasizes knowledge work and professional services as its core areas. We'll need to wait a few more weeks or months to see what CEO and co-founder Dario Amodei has planned. The company also prioritizes data security, intellectual property, and transparency in its AI development. Unlike OpenAI, which uses ChatGPT for tasks like writing scripts or sustainability essays, Anthropic aims to create solutions that corporate clients see as transformative before they commit.
A proxy war to establish the right narrative
Ultimately, the big question this intense competition raises is: where is all the funding going? It's well-known that training AI models is costly, and by the time a company optimizes a model to the point where answering a query costs just a dollar, a more advanced model might already be on the horizon.
While OpenAI offers its services for free, Anthropic has made it clear that it will not provide any of its offerings at no cost. How these models will evolve and how much Big Tech companies will invest in them, or develop them in-house, remains to be seen. These are questions that even OpenAI and Anthropic might not fully answer with their sophisticated chatbots.
How to Find Proxy Server Address??????♂?
In today's world, where protecting personal information and ensuring data security are top priorities, it is crucial for individuals with intermediate to advanced technical skills to understand how to set up and manage proxy servers.
Acting as an intermediary between your device and the internet, a proxy server offers protection, anonymity, and sometimes even enhances performance. To fully leverage these benefits, knowing your proxy server’s address is essential.
This guide will walk you through the details of finding server addresses across different browsers, mobile devices, and computer systems. Whether you're troubleshooting network issues or configuring a server, this guide will provide you with the technical knowledge you need.
What is a Proxy Server Address?
A proxy server address is essentially the IP address or hostname of a proxy server that processes your internet requests. For instance, it might be an IP address like 192.168.1.1 or a hostname such as proxy.example.com. When configured correctly, your device routes all of its internet traffic through this proxy server. The proxy server then forwards your requests to the appropriate websites or services and sends the responses back to your device.
The importance of knowing and using a proxy server address lies in several key benefits:
Finding a Proxy Server Address on Different Browsers
Each operating system—Windows, macOS, and Linux—offers methods to access and set up proxy settings. Here’s how you can locate a proxy server address on each platform:
Windows
macOS
Linux
These methods allow you to check and configure proxy settings on your device, whether through graphical interfaces or command line tools.
Welcome to "Three Truths and a Lie"! We'll serve you up with four headlines, but only one is fibbing. Can you spot the rogue headline among the truth-tellers? Let the guessing games begin!
Answer: ?? Oprah’s not actually about to turn AI into her personal clone. ?? Instead, it’s just a playful twist on the idea that Oprah’s insights are so profound, they could “teach” AI to be as amazing as she is. But rest assured, no AI robots are getting Oprah-fied anytime soon! ?????
Until next time, stay curious, stay tech-savvy, and we'll catch you in the next edition! ????
No boring stuff here – just tech with a side of swagger! ????
????? Champion of creativity & defender of authenticity, working to showcase the beauty & brilliance of humanity with art, design, animation, and more!
6 个月Glad to see more people and businesses taking legal action against illegal Ai data scraping to train generative platform. Generative plagiarized content that relies on stealing data should be seen a criminal, not innovative.