AI agents are designed to automatically complete tasks without users having to prompt them. Basically, agentic AI can execute actions without a user having to perform a specific triggering event first. Agentic AIs are a particular subtype of AI tools that focus on automating routine tasks.

New AWS group for agentic AI

According to the email, the new group for agentic artificial intelligence will be led by executive Swami Sivasubramanian, whose title on LinkedIn is VP of AWS Agentic AI. Sivasubramanian will report directly to AWS CEO Matt Garman, who wrote the internal email sent to AWS staff.

“Agentic AI has the potential to be the next multi-billion business for AWS,” Garman allegedly wrote in the email. He went on to say that, “We have the opportunity to help our customers innovate even faster and unlock more possibilities, and I firmly believe that AI agents are core to this next wave of innovation.”

Amazon’s recent push for agentic AI includes Alexa+

This new AWS group isn’t Amazon’s first foray into the world of agentic AI. Last week, the company announced that new agentic AI capabilities would soon be coming to Alexa, the company’s voice assistant AI tool.

The newly upgraded version will be called Alexa+, to distinguish it from the earlier versions. Alexa+ will reportedly be able to perform certain actions on its own even if users don’t issue a specific voice command or ask a question.

The Alexa+ service is free for Prime members but will cost $19.99 per month for non-members. Alexa+ will become available to U.S. households soon, starting with certain models of the HD smart display Echo Show devices. The first Echo Show models to get access to the agentic AI Alexa are 8, 10, 15, and 21.

More AWS reorganization in the works

The creation of the agentic AI group won’t be the only change at AWS. In another internal email sent on Tuesday, AWS senior vice president Peter DeSantis revealed several additional reorganizations are in the works. Three groups – hardware engineering, the Bedrock AI group, and the SageMaker AI group – will be moved underneath the compute organization. A new group will also be formed by combining customer experience and commerce.

The post ‘Next Multi-Billion Business for AWS’ Could Be Agentic AI appeared first on eWEEK.

The product lead for OpenAI’s video generator tool Sora said last Friday the company plans to incorporate the app into ChatGPT eventually. Rohan Sahai, the Sora product lead, shared this information during an “office hours” session on the Discord chat app, as reported by TechCrunch. Currently, Sora is only accessible through a standalone web app and is only available for paid ChatGPT users that have a Plus or Pro subscription.

No specific timeline for integrating Sora into ChatGPT

While Sahai said that OpenAI is actively working on integrating Sora into ChatGPT, he declined to offer a specific timeline as to when the video generator tool might become available on ChatGPT. However, he did indicate that the ChatGPT version of Sora may not provide the same kind of control and customization that is available through the standalone web app.

Previously, OpenAI mostly marketed the Sora AI model to artists, animators, videographers, and other creators that rely on the visual medium of film. The AI company’s announcement that Sora will be incorporated into ChatGPT signals that OpenAI is likely trying to broaden the appeal of the AI video generator and incentivize more users to subscribe to the paid plans.

What is the Sora AI video generator?

Sora is the video AI generator tool that was launched by OpenAI in early December 2024. The company first announced that Sora was in the works back in February 2024, but the product wasn’t ready to be released until the end of the year.

Currently, Sora is available to paid users who subscribe to ChatGPT Plus ($20/month) or ChatGPT Pro ($200/month). Plus users get videos of up to 5 seconds duration and 720p resolution, while Pro users get videos of up to 20 seconds duration and 1080p resolution. There is no free version of Sora available the same way there is a free version of ChatGPT.

On February 28, OpenAI announced on X that Sora is now available to Plus and Pro users in the EU, the U.K., Switzerland, Norway, Liechtenstein, and Iceland.

How does the Sora AI video generator work?

Sora videos can be generated by typing in a text prompt or by uploading an existing image or video. After submitting the initial artificial intelligence prompt, users can select the aspect ratio, resolution, duration, and the number of variations they wish to generate.

After the first AI video is generated, Sora users can edit the video in several ways. If they aren’t satisfied with the first version, they can remix it by describing changes and generating a new version. They can also recut the video to either trim or extend it as necessary. In addition, users can blend elements from two videos together or create an endless loop from a specific section of the video.

The post OpenAI Will Give ChatGPT Sora AI Video Gen, But When? appeared first on eWEEK.

iPhone users publicly document the mistake

Reports of the mistaken transcription first surfaced on Tuesday. iPhone users recorded videos of themselves saying the word “racist” aloud while the dictation feature was enabled. Apple’s AI model would initially write “Trump” before quickly correcting to “racist,” the actual word that was said.

After more and more iPhone users said they were also experiencing the issue, publications including The New York Times were able to replicate the error.

Apple blames swap on phonetic overlap

The issues appeared to begin after an update to Apple’s servers earlier in the week. Apple attributes the transcription error on a phonetic overlap between words that contain the r consonant, saying that it’s difficult for the dictation AI tool to distinguish between them.

“We are aware of an issue with the speech recognition model that powers Dictation and we are rolling out a fix today,” Apple told The Associated Press on Wednesday.

Apple said that other words that contain the r consonant could also cause the same bug. eWeek was not able to replicate the transcription error on Thursday, which suggests that the fix has indeed been rolled out to iPhones.

Former Siri team member casts doubt on Apple’s explanation

Not everyone believes Apple’s official statement, however. John Burkey, a former member of Apple Siri’s team who is still in “regular contact” with his former colleagues, told The New York Times that the bug seemed more like a deliberate prank than an accidental error, in his opinion.

Burkey said the fact that the word corrected itself indicates that a change was made to either the code or the data that powered the speech recognition AI tool. “This smells like a serious prank,” Burkey said. “The only question is: Did someone slip this into the data or slip into the code?”

The transcription error comes on the heels of Apple’s announcement earlier this week to invest more than $500 billion and hire more than 20,000 people in the United States. The tech company also announced plans to build a new factory in Texas, likely in response to Trump’s tariff on Chinese imports, which include iPhones and other Apple products.

The post Apple Blames AI Glitch After Dictation Replaces the Word ‘Racist’ With ‘Trump’ appeared first on eWEEK.

According to Reuters, which first broke the news, this marks the first antitrust lawsuit of its kind by a single company against Google’s AI Overviews feature.

Chegg filed its complaint on the same day it released the company’s fourth quarter and annual financial results for 2024. The education company’s total net revenue for the fourth quarter decreased 24% compared to the same period during the prior year. Annual total net revenues decreased 14% year over year. Because of these significant reductions, Chegg has hired Goldman Sachs to explore potential options, such going private or getting acquired.

Why Chegg is suing Google over AI

In his prepared remarks, Chegg CEO Nathan Schultz directly attributed these negative outcomes to Google’s search engine monopoly and use of artificial intelligence overviews. “We would not need to review strategic alternatives if Google hadn’t launched AI Overviews, or AIO, retaining traffic that historically had come to Chegg, materially impacting our acquisitions, revenue, and employees… Unfortunately, traffic is being blocked from ever coming to Chegg because of Google’s AIO and their use of Chegg’s content to keep visitors on their own platform,” he said.

Schultz went on to explain that with the generative AI overviews, Google has shifted from a “search engine” into an “answer engine.” He also laid out the three main arguments the complaint makes against Google:

• Reciprocal dealing: Google forces companies like Chegg to provide access to proprietary content in order to be included in its search engine.
• Monopoly maintenance: Google unfairly leverages its monopolistic hold over search engines to engage in anti-competitive conduct against companies like Chegg.
• Unjust investment: Google reaps the benefits of Chegg’s content without spending any money to produce it.

AI Overviews’ broader impact on publishers

While Chegg is the first individual company to take legal action, concerns over Google’s AI Overviews extend far beyond the education sector.

Last year, the News/Media Alliance, which represents over 2,000 publications, said that AI Overviews will have a “catastrophic” impact on web traffic and revenue.

A new study from brand management company Terakeet, published earlier on Tuesday, helped to quantify the impact that Google’s AI Overviews are having on search engine results. The study found that web pages included in AI Overviews had 3.2x as many clicks as pages that were excluded when it comes to top-ranked transaction queries.

The post Chegg Files Seminal Antitrust Lawsuit About Google’s AI Overviews Feature appeared first on eWEEK.

Musk’s personal history with OpenAI

Musk helped to cofound OpenAI back in 2015, alongside now-CEO Sam Altman and a group of other tech leaders. Musk left OpenAI in 2018 and later founded his own AI company, xAI, which is what Taylor was referencing in the board’s statement. Since leaving OpenAI, Musk has had a contentious relationship with both the company in general and Sam Altman specifically. His unsolicited buyout offer, and OpenAI’s subsequent rejection, are the latest development in what many view as an ongoing feud.

Buyout bid prompted by OpenAI restructuring

Musk’s buyout offer was somewhat prompted by a change in OpenAI’s corporate structure. The artificial intelligence company has an unusual structure, where the nonprofit entity OpenAI owns two for-profit corporate entities, OpenAI GP LLC and OpenAI Global LLC.

In September 2024, OpenAI announced that it planned to restructure into a for-profit benefit corporation that would no longer be controlled by the nonprofit board. The change would also allow Altman to receive equity for the first time.

Musk took issue with getting rid of OpenAI’s nonprofit status, saying “What they’re trying to do now is completely delete the nonprofit,” at the World Government Summit in Dubai on Feb. 13. Those remarks came a day after Musk released a court filing saying that he would not pursue the buyout offer as long as OpenAI continued to keep its original structure as a nonprofit charity.

Altman claps back at Musk’s buyout offer

Altman quickly countered Musk’s offer, even before the OpenAI board formally rejected the bid. The OpenAI CEO posted on X (formerly Twitter): “no thank you but we will buy twitter for $9.74 billion if you want.”

His post poked fun at Musk’s $44 billion purchase of Twitter, which he later rebranded as X. The app has declined precipitously in value since Musk acquired it and, as of December 2024, was worth about 72% less than Musk paid for it. Many people drew parallels between Musk’s takeover of Twitter and his attempted takeover of OpenAI.

Musk had his own brief retort for Altman. The X owner responded to Altman’s post with a single word: “Swindler.” Altman did not respond to Musk’s post, and the OpenAI board officially declined the offer on Feb. 14, 2025.

The post OpenAI Rejects Elon Musk’s $97.4 Billion Buyout Offer: “Not for Sale” appeared first on eWEEK.

Ellison says uploading national data is the “missing link”

In Ellison’s opinion, the U.S. won’t be able to take full advantage of artificial intelligence until all national data is uploaded into an AI model. He recommended consolidating fragmented databases into one huge population data set, then feeding it into an AI model — ideally, one created by his company — so it can be analyzed for improvement suggestions and cost-cutting ideas.

“We need to unify all the national data, put it into a database where it’s easily consumable by the AI model, and then ask whatever question you like,” he said. “That’s the missing link.”

Ellison particularly emphasized the potential for AI models to identify waste and fraud in the U.S. government. He also mentioned the so-called Department of Governmental Efficiency (DOGE), which is being run by fellow billionaire Elon Musk, who is also a friend of Ellison’s.

“We can provide high-quality services, save the government vast amounts of money, our populations will be healthier and the government will be spending less for better outcomes,” he said.

Critics say that making such a widespread, consolidated database could be the first step in creating a unified mass surveillance system that tracks the residents of the United States. Ellison expressed his support for such a surveillance system in September 2024, saying it would keep people “on their best behavior” on a call with Oracle financial analysts.

Oracle doubles down on AI technology

Ellison’s remarks came amidst Oracle’s huge and public push for artificial intelligence. Oracle is one of the founding partners of the Stargate initiative, alongside OpenAI, SoftBank, and MGX. The Stargate project aims to invest a total of $500 billion in U.S. AI infrastructure over the coming years. President Trump announced the project the day after assuming office.

While Oracle may not be the first AI company that people think of, the corporation is working to position itself as the go-to option for underlying infrastructure and networking architecture, primarily databases that can be used to underpin AI models. Ellison’s recent comments also indicate that the company hopes to become the foundation of any U.S. national AI models as well.

The post Ellison Thinks US Should Let Oracle AI Study All National Data appeared first on eWEEK.

One of the celebrities featured in the AI generated video, Scarlett Johansson, has now spoken out against the creation of AI deepfake videos. The actor’s voice and likeness have previously appeared in several other deepfakes without her permission.

Kanye West dropped by agency after series of scandals

Kanye West made headlines two weeks ago after his wife, Bianca Censori, appeared on the Grammys red carpet in a totally transparent dress. After the incident, West took to X to post a rant that included racist comments and antisemitic slurs. He called himself a Nazi and said that he thinks Hitler was “so fresh,” among other highly offensive comments.

He also aired an ad during the Super Bowl last Sunday that encouraged viewers to purchase from his website Yeezy.com, which only featured a swastika t-shirt. The website was taken down by Shopify for violating the terms of agreement. Now the site simply displays a blank white page with the message “Yeezy Stores Coming Soon.” His X account has since been deactivated as well.

On Tuesday, February 11, West was dropped by his talent agent, Daniel McCartney of 33 & West, via a public Instagram post. “Effective immediately, I’m no longer representing YE (F/K/A Kanye West) due to his harmful and hateful remarks that myself nor 33 & West can stand for,” McCartney wrote.

AI deepfake videos shows Jewish celebrities criticizing West

In the midst of these ongoing scandals, a video that purported to show many Jewish celebrities opposing West went viral. The video was a deepfake generated with artificial intelligence, and none of the celebrities actually participated in the making of it. Ori Bejerano and Guy Bar said they created this deepfake video, as reported by Business Insider, stating it was “… an artistic and cultural statement aimed at confronting rising antisemitism.”

The video showed Scarlett Johanssen in a white t-shirt emblazoned with a star of David in the middle of a hand lifting a finger. “Kanye” was written on the t-shirt below the graphic. The AI video generator also incorporated deepfakes of multiple other Jewish celebrities including Jerry Seinfeld and Adam Sandler.

Johanssen released a statement to PEOPLE magazine in reaction to the deepfake video going viral. In the statement, she said, “I am a Jewish woman who has no tolerance for antisemitism or hate speech of any kind. But I also firmly believe that the potential for hate speech multiplied by A.I. is a far greater threat than any one person who takes accountability for it. We must call out the misuse of A.I., no matter its messaging, or we risk losing a hold on reality.”

“I urge the U.S. government to make the passing of legislation limiting A.I. use a top priority; it is a bipartisan issue that enormously affects the immediate future of humanity at large,” the statement concluded.

The post Scarlett Johansson Calls Out Use of Her Likeness in AI Deepfake Video appeared first on eWEEK.

Macron’s deepfake video raises eyebrows

To promote the AI Action Summit, President Macron posted a video featuring AI-generated deepfakes of himself in various pop culture scenes. His AI doppelganger appears dancing in an ‘80s disco music video, making a cameo in the action series MacGyver, and even leading a YouTube hair tutorial.

In the video, Macron praised the deepfakes’ quality, saying “It’s pretty well done, it made me laugh.”

However, AI experts and critics, both in France and around the world, argue that the lighthearted approach could normalize deepfakes, a technology often used for misinformation, identify fraud, and even revenge porn. Given Macron’s status as president of one of the most prominent countries in Europe, many are concerned that his influence might sway people to dismiss the dangers of deepfakes.

“President Macron’s deepfake might seem like harmless fun to promote the AI Summit in Paris, but it is not in general a good thing,” Paul McKay, principal analyst at technology consultancy Forrester, told the BBC.

“Normalising deepfakes in this way should not be encouraged as it continues the difficulty with telling what is real and what isn’t, and is ultimately helping to establish what is fact from fiction,” he added.

View this post on Instagram

A post shared by Emmanuel Macron (@emmanuelmacron)

AI Summit kicks off French investment in AI

During the summit, President Macron positioned France’s €100+ billion AI investment as a strategic move to compete with the Stargate initiative recently announced by U.S. President Trump. Through this project, which was led by OpenAI, the U.S. will invest over $500 in artificial intelligence infrastructure. Macron’s announcement signals France’s ambition to remain a key player in the evolving AI landscape.

Several world leaders are attending the AI Summit, including India’s prime minister Narendra Modi, who is co-chairing the summit alongside Macron. Other notable political guests include U.S. vice-president JD Vance and Chinese vice-premier Zhang Guoqing. The summit runs through Wednesday, Feb. 11.

The post Macron’s Deepfake Stunt Sparks Debate as France Pledges €100 Billion Investment in AI appeared first on eWEEK.

When Mount Vesuvius erupted, its pyroclastic surges destroyed and preserved the Roman towns of Herculaneum and Pompeii. The blast also carbonized more than 1,000 scrolls, leaving them intact but too fragile to unroll.

The scrolls were discovered in a large villa in Herculaneum in the 1750s, and scholars and scientists have been trying to figure out a way to read the scrolls without destroying them ever since. This week, researchers announced a breakthrough thanks to a combination of modern technologies, including artificial intelligence and X-rays, which allows scientists to decipher the scrolls without unrolling them.

How AI helped decipher one of the scrolls

The scroll PHerc. 172 is one of three currently housed at the Bodleian Libraries at the University of Oxford in England; most of the other scrolls are held at the National Library of Naples in Italy.

PHerc. 172, a charred block of papyrus about the size of a candy bar, was first scanned at the Diamond Light Source Lab using a particle accelerator called a synchrotron. The synchrotron created scans of each papyrus layer, revealing visible ink. Researchers then used AI tools to assemble the images and enhance clarity of the written text.

According to the Oxford press release, the scholars applied a machine learning model focused solely on detecting ink, leaving interpretation to experts in ancient Greek.

“This lack of understanding [words] was, and remains, an advantage, as it ensures no speculative adjustments or interpretations are introduced, preserving the integrity of the original content,” the article said.

So far, scientists have identified a few Greek words, such as “disgust,” but much work remains to translate the entire scroll.

The Vesuvius Challenge

The breakthrough was made possible by the Vesuvius Challenge, launched in 2023 by several tech executives. The initiative offers awards for deciphering the scrolls using machine learning, computer vision, and geometry. Since its inception, the challenge has awarded $1.5 million worth of prizes to researchers, scientists, and scholars.

On Wednesday, the Vesuvius Challenge jointly announced the latest results in partnership with the Bodleian Libraries. The update on the Vesuvius Challenge website called it “incredible progress” while the Oxford press release hailed it a “historic breakthrough.”

The Vesuvius Challenge is calling on more scholars to accelerate the process.

The post AI Tools Help Decipher Scroll Burned In Mount Vesuvius Volcanic Eruption appeared first on eWEEK.

But another type of reader still accesses content for free: AI companies crawl the web relentlessly, scraping data to train their large language models (LLMs). Prince envisions a new internet utopia where this dynamic is flipped, so human readers can access content for free — but he wants AI companies to pay websites for scraping their data.

How AI models have broken the internet

Prince told Fortune that Google crawls each website for six times for each referred human user, which is already a lopsided ratio. But AI companies are even worse: OpenAI’s crawl-to-referral ratio is 250:1, and Anthropic’s is a whopping 6,000:1, according to Prince.

“For these new AI systems, the value of, ‘I’m going to take your data and then in exchange I’m going to send traffic back to your site’—that’s just going to break,” he explained. “And so we have to invent some other model.”

Prince says this dynamic disincentivizes internet content creators — and if creators stop making content, then AI tools will have less and less new data to train on, causing the models to degrade in quality over time.

Prince’s vision for an internet utopia

Prince has already begun rolling out a three-part plan at Cloudflare. Currently, site owners can opt into AI Audit, a collection of tools that allows them to see and control how often AI models use their site’s content. Prince says that AI Audit will become opt-out in the first half of 2025, so that AI crawlers will automatically be blocked.

After enough original content has been blocked, Prince says the power will return to content creators who decide how much to charge AI companies to use their content for training models. The final phase includes creating a marketplace for original content, but he doesn’t have that step fully figured out yet.

“It would be great if we got to a web that was back to: Humans get content for free, and bots pay a lot for that content,” Prince says.

The post Tech Billionaire’s Solution to Fixing the Internet Broken by AI? Utopia For Content Creators appeared first on eWEEK.