Microsoft Security Flaw Exposes 38TB Personal Data, More AI News

Microsoft AI researchers accidentally exposed 38 terabytes of private data due to a security flaw. However, the company has now taken corrective measures to fix this. The data leak was spotted on AI GitHub repository. The leak included backup data of two former employees' computers. It had keys, passwords and internal Teams chats. 

Microsoft has conducted a probe on the matter but found no evidence of unauthorised exposure. The company assured that customers do not need to do anything on their part. For future, the company has expanded its secret scanning service. For now, the repository is no longer accessible. 

UK's new AI principles prioritise transparency and access

The UK's antitrust regulator, the Competition and Markets Authority (CMA), has laid down seven principles of AI regulations for companies that use the technology. The CMA focused its attention on foundation models like OpenAI's GPT-4, Meta's Llama 2, and other LLMs. The seven principles aim to make artificial intelligence regulated by providing more accountability and authority to the companies. 

To avoid the risk of AI sector being dominated by a few companies who exercise market power, the CMA rules would be beneficial. The regulator has also asked for views and opinions from leading AI companies such as OpenAI, Google, Meta, Microsoft, and Nvidia, among others. 

AI-based camera helps spot tigers, sends alert with seconds

Artificial intelligence sensors are now being used to keep a track of wildlife. TraiGuard AI has developed cameras that can help spot tigers in the wild and send alerts to concerned authorities within seconds. Such technology could help with minimising the risks of attacks by wild cats, while also prevent poaching. 

The technology is currently under being tested in Madhya Pradesh. The TrailGuard AI has been previously tested in a reserve in East Africa. It is capable of sending alerts within 30 seconds of receiving data. Moreover, the battery is claimed to produce and send over 2,500 images on a single battery charge.

AI-reimagined Friends' reboot is nightmare for every fan of the series

If you are a fan of sitcoms, you must have heard about Friends, a show about six friends and their lives in Manhattan. Over the years, there have been several reboots and remakes of the popular TV series, with an Indian version as well. A new Reddit post has shared an AI reboot of Friends, which is giving nightmares to the fans of the series. 

The promo of this show begins with six friends dancing around a fountain of fire and water, followed by a bunch of monstrous creatures and animals. Moreover, the faces of the characters are also deformed. The AI version is created by Midjourney and Pika Labs, as the credit shows. 

WEF includes three new job categories in AI sector

Artificial intelligence sector has emerged as a vital part of the ever-evolving landscape of jobs. There have been several new job creations related to AI sector as the demand for the technology grows. Moreover, several big techs are now adding a new AI department to revolutionise the creativity. Adding to this, the World Economic Forum has added new job categories for AI — trainers, explainers, and sustainers. 

The trainers have the responsibility to craft the core of the AI, while explainers aim at making AI user-friendly. Sustainers, on the other hand, guarantee that the AI systems are being used effectively and responsibly. The WEF report also mentions that the AI sector is probably going to grow manifolds in the coming years, adding at least 39 percent more jobs than its current numbers.