AI Gets a Dose of Reality

Even artificial intelligence ('AI') suffers from growing pains... The New York Times is reportedly the latest in a sprawling list of entities to consider legal action against AI developer OpenAI, the creator of popular chatbot ChatGPT. Many language learning models like ChatGPT were fed an incredible sum of data scrubbed from the Internet, often from […] Not rendering correctly? View this e-mail as a web page [here](. [Empire Financial Daily] AI Gets a Dose of Reality By Kyle Wehrle --------------------------------------------------------------- [VIEWER DISCRETION ADVISED]( In a controversial new video from Whitney Tilson, he makes a bold prediction about the war in Ukraine. Due to the sensitive nature of this topic, viewer discretion is heavily advised. [Click here now to watch](. --------------------------------------------------------------- Even artificial intelligence ('AI') suffers from growing pains... The New York Times is reportedly the latest in a sprawling list of entities to consider legal action against AI developer OpenAI, the creator of popular chatbot ChatGPT. Many language learning models like ChatGPT were fed an incredible sum of data scrubbed from the Internet, often from illegitimate sites, in order to "learn" how to answer prompts correctly. It's unclear – but highly possible – that ChatGPT was trained on some degree of the Times' content. Naturally, the Times is up in arms about copyright issues. Up until recently, the Times and OpenAI had attempted to keep negotiations civil. There were talks of AI developers like OpenAI having to pay some sort of subscription-based licensing fee to use Times data to develop their technology. But now it seems media companies have caught on to the existential threat that AI poses to their business models. After all, why would somebody pay for a Times subscription when AI can increasingly strip that information from the original reporting of journalists, boil it down, and present it to you without you having to pay a single penny? News Corp (NWSA), the owner of the Wall Street Journal and the Times of London, has also been in heated discussions over IP with AI developers... Back in May, News Corp CEO Robert Thomson gave a damning opinion on AI data scrubbing and news summaries gathered by AI... These [AI news snippets] are super snippets, containing all the effort and insight of great journalism but designed so the reader will never visit a journalism website, thus fatally undermining that journalism. Thomson's fear is legitimate: tech titan Microsoft (MSFT), for example, has poured billions of dollars into OpenAI and has released an AI-integrated version of its search engine Bing called Bing Chat. However, it's highly unlikely that human journalists will be fully replaced anytime soon... Studies have shown that when AI is fed on too much AI-generated data, it "hallucinates." AI experts have dubbed the condition model autophagy disorder ("MAD"), which they liken to the historic inbreeding that led to serious health conditions among European royalty. Basically, when an AI model exhibits MAD, it confidently makes up facts. For this reason alone, we need human journalists somewhere in the daily news production cycle to generate human content for the AI model to keep functioning accurately. --------------------------------------------------------------- Recommended Link: ['The Prophet' issues a MAJOR warning...]( He called the exact day that Bitcoin peaked... He predicted the exact hour marijuana stocks would begin their collapse... He even predicted the exact day the market would bottom after the COVID collapse... He's nailed so many major market moves with such accuracy that CNBC called him "The Prophet"... Now, he's issuing the boldest warning of his career... right down to the hour. He's predicting an EXTREMELY rare event is set to occur... one that hasn't happened for 15 years. Millions of Americans will be taken by surprise when this happens. But you have an opportunity to prepare yourself NOW. [Click here to reveal his message in full](. --------------------------------------------------------------- Perhaps the more important question here is what will happen to AI developers and their language learning models and datasets as they face a tidal wave of lawsuits... Alphabet's (GOOGL) Google, for one, recently had a lawsuit filed against it that alleges the company "has been secretly stealing everything ever created and shared on the Internet by hundreds of millions of Americans" to train its chatbot Bard. Meanwhile, a slew of writers – including comedian Sarah Silverman – have accused OpenAI of knowingly and aggressively committing copyright infringement. Some other lawsuits pertain to instances of libel on ChatGPT's part, such as falsely attributing crimes to random people. Other cases go back a few years, like in 2020, when Thomson Reuters alleged that ROSS Intelligence copied the entirety of its Westlaw database – despite being denied a license – to use the database to train its competing generative AI-powered legal research platform. It's no wonder that tech firms – most notably Meta Platforms (META) – are increasingly deciding to keep the data they train their AI on a secret. AI is caught in a legal crossfire like we've never seen before... The worst-case scenario for AI is that AI developers will lose their copyright infringement lawsuits, pay burdensome fees to plaintiffs, and – most important – be forced destroy the datasets they've trained their AI chatbots on. If AI developers like OpenAI can survive such a deluge of legal fines, they would then need to recreate their AI datasets entirely from scratch using only permitted materials. As Vanderbilt University faculty member and AI researcher Daniel Gervais once told NPR, "Copyright law is a sword that's going to hang over the heads of AI companies for several years unless they figure out how to negotiate a solution." It's worth noting here that it isn't just journalists, creatives, and consumers up in arms about how AI is bypassing their creative copyrights... AI-oriented companies are also suing each another. For example, Getty Images is currently accusing Stability AI of stealing 12 million copyrighted images to create a rival platform. (The icing on the cake here is that Stability AI's image generator often produce images that feature distorted versions of Getty Images' watermark!) But we're already getting an idea of where the law is going with AI... Earlier this month, a federal judge upheld an argument from the U.S. Copyright Office that "artworks" generated by AI models aren't covered by copyright protection. The decision likely comes as a breath of fresh air for the many writers taking part in the Writers Guild of America's strike for labor safeguards against AI. But for AI developers – and companies looking to get rich off easily generated AI content – the picture is looking less rosy than it did a few months ago when ChatGPT was first unveiled. The reality is that the goals and aspirations of AI companies will need to change in the face of developing legal frameworks if they want to protect their datasets. What once seemed like an unbridled cash grab on the horizon could soon become a path ridden with legal potholes and licensing fees if AI developers are caught scanning copyrighted data without licensing. The key detail to watch here is which companies respond with the best new roadmaps to get the most of AI without running afoul of new laws... These would be the ones poised to win the future of AI. Regards, Kyle Wehrle August 29, 2023 Editor's note: According to Empire Financial Research CEO Whitney Tilson, AI – which could represent an astounding $87 trillion market by 2030 – has the potential to create the world's first trillionaires. Don't be left behind as this wave of wealth unfolds – find out how Whitney says to take advantage of this massive shift [right here](. --------------------------------------------------------------- If someone forwarded you this e-mail and you would like to be added to the Empire Financial Daily e-mail list to receive e-mails like this every weekday, simply [sign up here](. © 2023 Empire Financial Research. All rights reserved. Any reproduction, copying, or redistribution, in whole or in part, is prohibited without written permission from Empire Financial Research, 1125 N. Charles Street, Baltimore, Maryland 21201 [www.empirefinancialresearch.com.]( You received this e-mail because you are subscribed to Empire Financial Daily. [Unsubscribe from all future e-mails](