It’s not just “violence,” social media also makes kids crazy?
Schools Sue Social-Media Platforms Over Alleged Harms to Students
Plaintiffs’ lawyers are pitching school boards throughout the country to file lawsuits against social-media companies on allegations that their apps cause classroom disciplinary problems and mental-health issues, diverting resources from education.
Nearly 200 school districts so far have joined the litigation against the parent companies of Facebook, TikTok, Snapchat and YouTube. The suits have been consolidated in the U.S. District Court in Oakland, Calif., along with hundreds of suits by families alleging harms to their children from social media.
The lawsuits face a test later this year when a judge is expected to consider a motion by the tech companies to dismiss the cases on grounds that the conduct allegedly causing the harm is protected under the internet liability shield known as Section 230.
This raises the question: Is this the best method for training AI? How can anyone vet this data?
https://www.bespacific.com/a-i-brings-shadow-libraries-into-the-spotlight/
A.I. brings shadow libraries into the spotlight
The New York Times [free link ] – to see this text scroll down the page: ” Large language models. or L.L.M.s, the artificial intelligence systems that power tools like ChatGPT, are developed using enormous libraries of text. Books are considered especially useful training material, because they’re lengthy and (hopefully) well-written. But authors are starting to push back against their work being used this way. This week, more than 9,000 authors, including Margaret Atwood and James Patterson, called on tech executives to stop training their tools on writers’ work without compensation. That campaign has cast a spotlight on an arcane part of the internet: so-called shadow libraries, like Library Genesis, Z-Library or Bibliotik, that are obscure repositories storing millions of titles, in many cases without permission — and are often used as A.I. training data. A.I. companies have acknowledged in research papers that they rely on shadow libraries. OpenAI’s GPT-1 was trained on BookCorpus, which has over 7,000 unpublished titles scraped from the self-publishing platform Smashwords. To train GPT-3, OpenAI said that about 16 percent of the data it used came from two “internet-based books corpora” that it called “Books1” and “Books2.” According to a lawsuit by the comedian Sarah Silverman and two other authors against OpenAI, Books2 is most likely a “flagrantly illegal” shadow library. These sites have been under scrutiny for some time. The Authors Guild, which organized the authors’ open letter to tech executives, cited studies in 2016 and 2017 that suggested text piracy depressed legitimate book sales by as much as 14 percent. Efforts to shut down these sites have floundered. Last year, the F.B.I., with help from the Authors Guild, charged two people accused of running Z-Library with copyright infringement, fraud and money laundering. But afterward, some of these sites were moved to the dark web and torrent sites, making it harder to trace them. And because many of these sites are run outside the United States and anonymously, actually punishing the operators is a tall task.”
Always interesting, never amusing.
IBM Report: Half of Breached Organizations Unwilling to Increase Security Spend Despite Soaring Breach Costs
AI/Automation cut breach lifecycles by 108 days; $470,000 in extra costs for ransomware victims that avoid law enforcement; Only one third-of organizations detected the breach themselves
IBM Security today released its annual Cost of a Data Breach Report,1 showing the global average cost of a data breach reached $4.45 million in 2023 – an all-time high for the report and a 15% increase over the last 3 years. Detection and escalation costs jumped 42% over this same time frame, representing the highest portion of breach costs, and indicating a shift towards more complex breach investigations.
According to the 2023 IBM report, businesses are divided in how they plan to handle the increasing cost and frequency of data breaches. The study found that while 95% of studied organizations have experienced more than one breach, breached organizations were more likely to pass incident costs onto consumers (57%) than to increase security investments (51%).
No comments:
Post a Comment