AI Research

“Unknown actor” using AI to impersonate Rubio, State Department cable shows

Published

2 months ago

July 8, 2025

An “unknown actor” created a Signal account and used artificial intelligence to impersonate Secretary of State Marco Rubio and contact government and foreign officials, according to a State Department cable obtained by CBS News.

The cable was sent to “All Diplomatic and Consular Posts,” and alerted them that the account was created in mid-June 2025 with the display name marco.rubio@state.gov.

“The individual contacted at least five non-Department individuals, including three foreign ministers, a U.S. governor, and a U.S. member of Congress,” the cable said. “The actor left voicemails on Signal for at least two targeted individuals, and in one instance, sent a text message inviting the individual to communicate on Signal.”

In a statement, the State Department said it is “aware of this incident and is currently investigating this matter.” The Department said it “continuously takes steps to improve the department’s cybersecurity posture to prevent future incidents.”

The story was first reported by The Washington Post.

The cable says the alleged imposter “likely aimed to manipulate targeted individuals using AI-generated text and voice messages, with the goal of gaining access to information or accounts.”

According to the cable, this effort resembled an incident in May 2025 that was investigated by the FBI of someone impersonating other senior U.S. government officials.

There was a separate and unrelated campaign that began in April 2025 of Russia-linked cyber threat actor conducting a spear-fishing campaign that targeted personal Gmail accounts associated with think tank scholars, Eastern European-based activists and dissidents, journalists and former State Department officials. In that incident, the actor allegedly posed as a fictitious State Department official and invited the targets to a meeting in an attempt to link a third-party application to their Gmail accounts. The cable said that if clicked, the application would “almost certainly” grant the actor persistent access to the Gmail accounts’ contents. For this second campaign, industry partners attributed it to a cyber threat actor associated with the Russian Foreign Intelligence Service (SVR).

The cable also listed other instances of threat actors impersonating Department accounts in the past few years.

The news also follows “Signalgate” in March, when journalist Jeffrey Goldberg was inadvertently added to a group chat on Signal that included Rubio, Vice President JD Vance, Defense Secretary Pete Hegseth and then-National Security Adviser Mike Waltz. On the chat, the officials discussed plans to strike Houthi targets in Yemen and Waltz was later ousted due in part to the incident.

Source link

AI Research

USA TODAY rolls out AI answer engine to all users

Published

10 minutes ago

September 15, 2025

, USA TODAY

Gannett, USA TODAY’s parent company, has fully implemented generative AI engine DeeperDive for USA TODAY’s audience of more than 195 million monthly unique visitors.

DeeperDive uses the high-quality content created by reporters and editors of the USA TODAY Network to deliver clear, timely GenAI conversations to readers. The technology was created by Taboola. Gannett is the first U.S. publisher to fully embed the AI-answer engine.

The step aligns with the company’s commitment to embrace innovation for the benefit of its readers, Michael Reed, chairman and CEO of Gannett, said in a statement.

“The Taboola partnership gives us the opportunity to further deliver on our promise to enrich and empower the communities we serve because DeeperDive provides our valued audiences with trusted relevant content,” Reed said.

Because it sources its responses solely from trusted USA TODAY and USA TODAY Network journalism and content, DeeperDive interacts with readers to deliver a sharper understanding of the topics users want to know about.

Other highlights include more curated advertising, Reed said. A DeeperDive beta was launched in June to a percentage of readers and was expanded after initial performance exceeded expectations.

DeeperDive’s technology spans various coverage areas, answering reader questions about travel, their local communities, sports, political updates and more.

In the next phase of the collaboration, AI agents will be tested to give readers access to seamless, easy purchasing options tailored to their specific needs and interests, Reed said.

Adam Singolda, CEO and founder of Taboola, called the partnership with Gannett a “once-in-a-generation” opportunity.

“With DeeperDive, we’re moving the industry from page views to Generative AI conversations, and from clicks to transactions rooted in what I see as the most valuable part of the LLM market – decisions that matter,” Singolda said in a statement. LLM refers to large language models, like ChatGPT.

“Consumers may ask questions using consumer GenAI engines, but when it comes to choices that require trust and conviction, where to travel with their family, which financial step to take, or whether to buy a product – USA TODAY is where they turn,” added Singolda.

Source link

AI Research

OpenAI makes $300 billion gamble on Oracle computing power to expand artificial intelligence capacity

Published

35 minutes ago

September 15, 2025

Wayne Williams

OpenAI signs $300 billion Oracle contract starting in 2027 to expand AI capacity
Oracle shares jump over 40 percent after reporting $317 billion in future revenue
Deal raises risks as OpenAI loses money and Oracle takes on heavy debt

OpenAI has signed a contract with Oracle to buy $300 billion worth of computing power over the next five years, according to the Wall Street Journal.

This makes it one of the largest cloud deals ever struck.

The contract will begin in 2027 and is expected to reshape how OpenAI builds and runs its artificial intelligence models.

A huge gamble

The agreement will require 4.5 gigawatts of power capacity, which is enough electricity to supply about four million homes.

It shows how the rush to build AI data centers is driving new highs in technology spending even as questions remain over whether demand will justify such commitments.

Oracle disclosed in its latest earnings report that it added $317 billion in future contract revenue during the quarter ending August 31, partly due to the OpenAI deal.

The news sent Oracle shares soaring by more than 40 percent in a single day. That surge increased Oracle Chairman Larry Ellison’s wealth by more than $100 billion, and saw him overtake Elon Musk as the world’s richest person with a net worth close to $400 billion.

The deal is not without massive risk for both parties, however.

For OpenAI, the agreement provides a new source of computing power after years of relying exclusively on Microsoft’s Azure cloud, but WSJ says the company, which reported about $10 billion in revenue this year, will owe Oracle an average of $60 billion annually under the agreement.

The startup is losing money and has told investors it does not expect to turn a profit until 2029.

Oracle, meanwhile, will have to borrow heavily to finance the AI chips and infrastructure needed to deliver the contract.

Plus, as WSJ reported, “The deal rests on the assumption ChatGPT will continue its explosive growth and be adopted by billions of people across the world, as well as major businesses and governments.”

Industry analysts say the partnership underlines both the promise and the strain of the AI boom. Spending on chips, servers, and data centers worldwide is projected to reach $2.9 trillion by 2028.

Whether OpenAI’s growth can keep pace with its commitments remains an open question.

Source link

AI Research

Artificial Intelligence Technology Solutions Inc. Announces Commercial Availability of Radcam Enterprise

Published

59 minutes ago

September 15, 2025

S&P Capital IQ

Artificial Intelligence Technology Solutions Inc. along with its subsidiary, Robotic Assistance Devices Inc. (RAD-I), announced the commercial availability of RADCam? Enterprise, a proactive video security platform now compatible with the industry’s leading Video Management Systems (VMS). The intelligent talking camera can be integrated quickly and seamlessly into virtually any professional-grade video system.

The Company first introduced the RADCam Enterprise initiative on May 5, 2025, highlighting its expansion beyond residential applications into small medium business (SMB) and enterprise markets. With today’s availability, RAD-I will deliver the solution through an untapped niche in the security industry, specifically security system integrators and security system distributors. RADCam Enterprise brings an intelligent “operator in the box” capability, enabling immediate talk-down to potential threats before human intervention is required.

The device integrates a speaker, microphone, and high-intensity lighting, allowing it not only to record but also to actively engage. At the same time, the solution is expected to deliver gross margins consistent with the Company’s established benchmarks. RADCam Enterprise distinguishes itself from the original residential version of RADCam by integrating RAD’s agentic AI platform, SARA (Speaking Autonomous Responsive Agent) as well as being compatible with RADSoC and industry leading Video Management Systems. RADCam Enterprise is available immediately through RAD-I’s network of channel partners and distributors.

Pre-orders are open at giving clients the opportunity to be among the first to deploy the solution. Designed for broad use across industries including logistics, retail, education, and commercial real estate, RADCam Enterprise provides clients and integrators with new ways to modernize security operations using proven AI-driven tools. RAD delivers these cost savings via a suite of stationary and mobile robotic solutions that complement, and at times, directly replace the need for human personnel in environments better suited for machines.

All RAD technologies, AI-based analytics and software platforms are developed in-house. The Company’s operations and internal controls have been validated through successful completion of its SOC 2 Type 2 audit, which is a formal, independent audit that evaluates a service organization’s internal controls for handling customer data and determines if the controls are not only designed properly but also operating effectively to protect customer data. Each Fortune 500 client has the potential of making numerous orders over time.

AITX is an innovator in the delivery of artificial intelligence-based solutions that empower organizations to gain new insight, solve complex challenges and fuel new business ideas. Through its next-generation robotic product offerings, AITX’s RAD, RAD-R, RAD-M and RAD-G companies help organizations streamline operations, increase ROI, and strengthen business. The Company has no obligation to provide the recipient with additional updated information.

No information in this publication should be interpreted as any indication whatsoever of the Company’s future revenues, results of operations, or stock price.

Source link