Technical data

Criminals use fake AI voice to swindle UAE bank over $ 35 million • The Register


In short Authorities in the United Arab Emirates have requested the assistance of the US Department of Justice to investigate a case involving a bank manager who was defrauded by transferring $ 35 million to criminals by someone using a fake voice generated by AI.

The employee received a call to move funds belonging to the company by someone claiming to be a director of the company. He had also previously seen emails showing that the company was planning to use the money for an acquisition and had hired a lawyer to coordinate the process. When the bogus manager asked him to transfer the money, he did so thinking it was a legitimate request.

But it was all a scam, according to US court documents reported by Forbes. The criminals used “deep voice technology to simulate the director’s voice,” he said. Now officials in the United Arab Emirates have asked the DoJ to hand over details of two US bank accounts, where more than $ 400,000 of the stolen money was deposited.

Investigators believe there are at least 17 people involved in the heist.

AI systems must see the human perspective

Facebook has partnered with 13 universities from nine countries to compile Ego4D, a dataset containing over 2,200 hours of first-person video, where 700 participants were filmed performing daily activities like cooking or cooking. to play video games.

The anti-social network hopes that Ego4D will unlock new capabilities in augmented and virtual reality or robotics. New models formed on this data can be tested on a range of tasks, including episodic memory, predicting what happens next, coordinating hand movements to manipulate objects, and social interaction.

“Imagine your AR device displaying exactly how to hold sticks during a drum lesson, guiding you through a recipe, helping you find your lost keys, or recalling memories as holograms come to life in front of you,” Facebook said. in a blog Post.

“Next-generation AI systems will need to learn lessons from an entirely different kind of data – videos that show the world at the center of the action, rather than on the fringes,” added Kristen Grauman, senior researcher at Facebook.

Researchers will have access to Ego4D later next month subject to a data use agreement.

Microsoft Translator AI Software

Microsoft Translator, language translation software powered by neural networks, can now translate over 100 different languages.

Twelve new languages ​​and dialects were added to Microsoft Translator this week, including: from endangered languages ​​like Bashkir spoken by a Turkish Kiptchak ethnic group originally from Russia to more common jargons like Mongolian. Microsoft Translator now supports 103 languages.

“One hundred languages ​​is a good step for us to reach our ambition that everyone can communicate no matter what language they speak,” said Xuedong Huang, technical researcher at Microsoft and CTO of Azure AI.

Xuedong said the software is based on a multilingual AI model called Z-code. The system deals with text and is part of Microsoft’s efforts to create a larger multimodal system capable of handling images, text, and audio dubbed the XYZ Code Vision. Microsoft Translator is deployed in a range of services, including the Bing search engine and offered as an API on its Azure Cognitive Services cloud platform.

ShotSpotter sues Vice for libel, claims $ 300 million in damages

The controversial AI gun detection company Shotspotter sued Vice, claiming that its business had been unfairly tarnished by a series of articles published by the outlet.

“On July 26, 2021, Vice launched a defamatory campaign in which he falsely accused ShotSpotter of conspiring with the police to fabricate and alter evidence to charge black men with crimes they did not commit,” says the complaint.

ShotSpotter accused the publication of inaccurately portraying the company’s technology and actions to “cultivate a ‘subversive’ brand” used to sell products advertised in its “sponsored content.”

The company made headlines when evidence used to try to prove that a black man shot and killed another man in a trial was withdrawn. Defense attorney accused ShotSpotter employees of tampering with evidence to support the police case. Vice allegedly made false claims that the company routinely used its software to label loud sounds as gunshots in order to help law enforcement prosecute innocent suspects in shootings.

When Vice reporters received proof that this was not the case, they refused to correct their factual inaccuracies, according to the lawsuit. ShotSpotter argued that the articles had ruined his reputation and that he now wants Vice to pay $ 300 million in damages.

State of AI 2021

The annual State of AI report has come out, compiled by two UK tech investors, recapping AI trends and developments this year.

The fourth report by Nathan Benaich, VC at Air Street Capital, and Ian Hogarth, co-founder of music app Songkick and angel investor, focuses on transformers, a type of machine learning architecture known to power models language giants like OpenAI’s GPT. -3 or Google’s BERT.

Transformers aren’t just useful for generating text; they have proven themselves in other fields, such as computer vision or biology. Machine learning technology also continues to mature – developers are deploying more systems to solve real-world problems, such as optimizing energy through national power grids or warehouse logistics for supermarkets.

This also applies to military applications, the couple warned. “AI researchers have traditionally viewed the AI ​​arms race as a figurative race – simulated air battles between competing AI systems performed in laboratories – but this is changing with reports of recent weapon use. autonomous by various armies. “

You can read the full report here. ®


Leave a Reply

Your email address will not be published.