Google is facing a class-action lawsuit accusing the company of unauthorised data scraping from millions of users through its AI tools. The lawsuit alleges that Google violated copyright laws by using the data to train and develop its AI products, reports CNN. Filed by Clarkson Law Firm against Google, its parent company Alphabet, and AI subsidiary DeepMind, the lawsuit claims that Google has been covertly collecting and appropriating vast amounts of internet content from hundreds of millions of Americans.
Similar to a previous lawsuit against OpenAI, the firm behind ChatGPT, the lawsuit highlights concerns over the misuse of publicly available data to train AI models.
ALSO READ: Microsoft implements new job cuts, laying off 276 employees
According to IANS, the complaint argues that Google's actions encompassed the entirety of users' digital footprint, including their creative and copyrighted works, all of which were utilized to enhance the company's AI offerings.
As of now, Google has not responded to the lawsuit. Notably, last week, Google updated its privacy policy, clarifying that publicly available data can be utilised to train its AI models. The revised policy emphasises that any public content posted online could be used for training Bard, future iterations, and other generative AI products.
ALSO READ: Amazon Prime Day Sale safety: Avoid these common scams and protect your money
Reportedly, The usage of publicly available information by companies for training large language models has drawn criticism from various quarters. The concern over data privacy and the use of personal information for AI development remains a contentious issue.
Furthermore, the tech giant is currently testing an artificial intelligence (AI) program designed to answer medical questions accurately, according to reports. The AI tool, known as Med-PaLM 2 (a variant of PaLM 2), has been undergoing testing at the Mayo Clinic and other research hospitals since April. Med-PaLM 2 is built upon the language model called Bard, developed by Google.
Inputs from IANS
