Latest news with #Grok
Egypt Independent
11 minutes ago
- Business
- Egypt Independent
Elon Musk isn't happy with his AI chatbot. Experts worry he's trying to make Grok 4 in his image
Musk was not pleased. 'Major fail, as this is objectively false. Grok is parroting legacy media,' Musk wrote, even though Grok cited data from government sources such as the Department of Homeland Security. Within three days, Musk promised to deliver a major Grok update that would 'rewrite the entire corpus of human knowledge,' calling on X users to send in 'divisive facts' that are 'politically incorrect, but nonetheless factually true' to help train the model. 'Far too much garbage in any foundation model trained on uncorrected data,' he wrote. On Friday, Musk announced the new model, called Grok 4, will be released just after July 4th. The exchanges, and others like it, raises concerns that the world's richest man may be trying to influence Grok to follow his own worldview – potentially leading to more errors and glitches, and surfacing important questions about bias, according to experts. AI is expected to shape the way people work, communicate and find information, and it's already impacting areas such as software development, healthcare and education. And the decisions that powerful figures like Musk make about the technology's development could be critical. Especially considering Grok is integrated into one of the world's most popular social networks – and one where the old guardrails around the spread of misinformation have been removed. While Grok may not be as popular as OpenAI's ChatGPT, its inclusion in Musk's social media platform X has put it in front of a massive digital audience. 'This is really the beginning of a long fight that is going to play out over the course of many years about whether AI systems should be required to produce factual information, or whether their makers can just simply tip the scales in the favor of their political preferences if they want to,' said David Evan Harris, an AI researcher and lecturer at UC Berkeley who previously worked on Meta's Responsible AI team. A source familiar with the situation told CNN that Musk's advisers have told him Grok 'can't just be molded' into his own point of view, and that he understands that. xAI did not respond to a request for comment. Concerns about Grok following Musk's views For months, users have questioned whether Musk has been tipping Grok to reflect his worldview. In May, the chatbot randomly brought up claims of a white genocide in South Africa in responses to completely unrelated queries. In some responses, Grok said it was 'instructed to accept as real white genocide in South Africa'. Musk was born and raised in South Africa and has a history of arguing that a 'white genocide' has been committed in the nation. A few days later, xAI said an 'unauthorized modification' in the extremely early morning hours Pacific time pushed the AI chatbot to 'provide a specific response on a political topic' that violates xAI's policies. As Musk directs his team to retrain Grok, others in the AI large language model space like Cohere co-founder Nick Frosst believe Musk is trying to create a model that pushes his own viewpoints. 'He's trying to make a model that reflects the things he believes. That will certainly make it a worse model for users, unless they happen to believe everything he believes and only care about it parroting those things,' Frosst said. What it would take to re-train Grok It's common for AI companies like OpenAI, Meta and Google to constantly update their models to improve performance, according to Frosst. But retraining a model from scratch to 'remove all the things (Musk) doesn't like' would take a lot of time and money – not to mention degrade the user experience – Frosst said. 'And that would make it almost certainly worse,' Frosst said. 'Because it would be removing a lot of data and adding in a bias.' A Grok account on X is displayed on a phone screen. Jakub Porzycki/NurPhoto/Shutterstock Another way to change a model's behavior without completely retraining it is to insert prompts and adjust what are called weights within the model's code. This process could be faster than totally retraining the model since it retains its existing knowledge base. Prompting would entail instructing a model to respond to certain queries in a specific way, whereas weights influence an AI model's decision-making process. Dan Neely, CEO of Vermillio which helps protect celebrities from AI-generated deepfakes, told CNN that xAI could adjust Grok's weights and data labels in specific areas and topics. 'They will use the weights and labeling they have previously in the places that they are seeing (as) kind of problem areas,' Neely said. 'They will simply go into doing greater level of detail around those specific areas.' Musk didn't detail the changes coming in Grok 4, but did say it will use a 'specialized coding model.' Bias in AI Musk has said his AI chatbot will be 'maximally truth seeking,' but all AI models have some bias baked in because they are influenced by humans who make choices about what goes into the training data. 'AI doesn't have all the data that it should have. When given all the data, it should ultimately be able to give a representation of what's happening,' Neely said. 'However, lots of the content that exists on the internet already has a certain bent, whether you agree with it or not.' It's possible that in the future, people will choose their AI assistant based on its worldview. But Frosst said he believes AI assistants known to have a particular perspective will be less popular and useful. 'For the most part, people don't go to a language model to have ideology repeated back to them, that doesn't really add value,' he said. 'You go to a language model to get it to do with do something for you, do a task for you.' Ultimately, Neely said he believes authoritative sources will end up rising back to the top as people seek places they can trust. But 'the journey to get there is very painful, very confusing,' Neely said and 'arguably, has some threats to democracy.'
Business Insider
4 hours ago
- Politics
- Business Insider
Your member of Congress might be using ChatGPT
In December, Rep. Thomas Massie used an analogy for foreign aid that was an instant hit among his libertarian and America-First Republican fans. "US foreign aid spending is like watering the neighbor's yard while your house is on fire," the Kentucky Republican posted on X, adding a fire emoji. Fox News wrote an article about it, and two months later, the libertarian student group "Young Americans for Liberty" turned it into an Instagram post. As it turns out, Massie didn't come up with the line himself. Grok did. Massie told BI this month that he ripped the phrase from a speech he asked the xAI-developed chatbot to generate using his voice. He said he's done this more than once. "Out of five paragraphs, I'll find one sentence that's good," Massie said. "But it makes it worth doing." Leaning on AI for speechwriting is an apparently bipartisan affair on Capitol Hill. "I'll type in some phrases and say, can we make this more punchy?" Democratic Rep. Ro Khanna of California told BI, adding that he began using ChatGPT "almost like an editor" in the last year. "There was some speech I gave where it edited in a couple of lines that people thought, 'Wow, that's really good,'" Khanna said. Congress has developed a reputation for lagging behind the public when it comes to adopting new technology. Plenty of lawmakers told BI that they have yet to get into using AI, either because they're skeptical that it will be useful for them or they just haven't gotten around to it. But several lawmakers have begun to casually adopt the technology, most often as a search engine and research tool. Khanna said he uses both ChatGPT and Grok, turning to the technology "two to three times per day." Massie, who uses Grok because of its convenient placement within the X app, said he uses the chatbot for "anything." 'Impressively good at certain things and pretty miserable at some things' As Sen. Ron Johnson of Wisconsin has waged a fight to make deeper cuts to federal spending as part of the "Big Beautiful Bill," he's been consulting with Grok. "I got up at 3 o'clock in the morning with an idea to use it," the Wisconsin Republican told BI in early June. He said the technology's been useful for running the numbers on the bill's impact on the deficit and to find documents that support his arguments. "It's really great at identifying sources without me having to crawl around in government forms." In some ways, members of Congress are just doing what other Americans are doing. More and more people are using AI at work, according to a recent Gallup poll, with 40% of employees saying they use it a few times per year. Another 19% say they use it frequently, while 8% say they use it on a daily basis. Sen. Ted Cruz of Texas, a champion of a controversial provision in the "Big Beautiful Bill" that would restrict state's ability to regulate AI for 10 years, told BI that while he "would not claim to be a sophisticated AI user," he's been using ChatGPT as an "enhanced search engine." Cruz said he recently asked an AI chatbot about his own record, when he "could not remember when I had first taken a public position" on a particular policy area. "It gave a very thorough answer, going back to an interview I'd done in 2012 and a comment I'd made in 2014," Cruz said. "That research previously would have required some staff assistance, spending hours and hours, and you still wouldn't have found anything." Large language models like ChatGPT and Grok are known to sometimes present false information as fact — known as "hallucinating." For Democratic Sen. Elizabeth Warren of Massachusetts, that's enough to discourage her from using it. "It lies," Warren told BI. "I've tried using it, and it gets things wrong that I already know the answer to. So when I see that, I've lost all confidence." Democratic Sen. Chris Murphy of Connecticut said he's tried ChatGPT and has been disappointed by its apparent limitations, even when carrying out more basic tasks. In one instance, Murphy said he asked ChatGPT to generate a list of his Democratic colleagues ordered alphabetically by first name, only for it to include retired senators. "It seems to be impressively good at certain things and pretty miserable at some things," Murphy said. Even those who are otherwise fans of the technology said they're aware that they could be getting fed incorrect information. "My chief of staff has astutely warned me that AI is often confidently wrong," Johnson said. "So you really have to be careful in how you phrase your questions." "It definitely hallucinates on you," Massie said. "It told me there was a Total Wine and More in Ashland, Kentucky, and no such thing exists."
Hindustan Times
11 hours ago
- Politics
- Hindustan Times
Large barricades erected around White House and Treasury? Video sparks speculation
A video circulating on social media shows construction crews erecting large anti-scale barricades around the White House and the US Treasury building. This has prompted widespread speculation about the reason behind the sudden security measures. A video showing construction crews erecting barricades around the White House has surfaced. (REUTERS) While some users suggested the barricades could be related to preparations for upcoming July 4 celebrations, others speculated they might be tied to potential protests or political developments. One person commented, 'I believe they're getting ready for the new Kings protest on July 3 or fourth.' Another user added, 'Yeah... when they break out the anti-scale fencing at night, it's never because they're prepping for a bake sale. That's usually code for: Some kind of announcement is coming, possibly with serious public blowback. A trial verdict, indictment, or legal action.' A third person commented, '4th of July protests, planned in DC, likely to be a riot. Don't leave your bricks out.' Another user wrote, 'I never saw the White House barricaded for the 4th of July parade or celebration in the 3 years I was stationed there from 2021-2024. Very odd.' Another commented, 'I do recall seeing a post somewhere talking about a No Kings 2.0 Day on July 4th. This is likely in response to that among many similar things.' Adding to the speculation, AI chatbot Grok responded to a user by citing a planned protest: 'There is a planned 'National March on Washington' protest against U.S. military action in Iran on June 28, 2025, likely prompting the barricades around the White House and Treasury. The event, organized by the ANSWER Coalition, is expected to draw large crowds. No official statements confirm this, but the timing and anti-scale barriers suggest crowd control for this protest. Other possibilities, like unannounced events or routine security, lack evidence.' Also Read: Is White House on fire today? Debunking viral post At this time, the video remains unverified, and Hindustan Times has not independently confirmed its authenticity. The White House has not issued a statement in response to the footage.
Yahoo
17 hours ago
- Business
- Yahoo
Oracle Benefits From AI Cloud Service Adoption: A Sign of More Upside?
Oracle's ORCL collaboration with xAI to deploy its Grok models via the Oracle Cloud Infrastructure ('OCI') is expected to further drive momentum in the company's cloud services and license support revenues in the near term. OCI spans domains like compute, databases and AI services, providing end-to-end solutions for enterprise workloads. Oracle's 23 AI database has helped businesses integrate AI solutions by automating workflows and providing flexibility to tailor models to business-specific needs. Oracle had previously demonstrated its ability to host, train and scale various models through partnerships with Cohere, LLAMA 2 and NVIDIA AI Enterprise. The latest partnership with xAI is expected to significantly increase OCI compute, storage and network usage. Additionally, the seamless deployment of Grok models is likely to encourage long-term contract renewals. Looking forward, total cloud revenues are expected to grow 26-30% for the first quarter of 2026, and by over 40% for fiscal 2026. The company also noted that for fiscal 2026, cloud infrastructure revenues are projected to grow even more than 70%, up from 51% in the prior year. Oracle faces tough competition from players like Alphabet GOOGL and Amazon AMZN. Alphabet's Google Cloud invests heavily in custom AI chips called TPUs and recently introduced its latest version, Ironwood. The company recently launched the 'Cloud Wide Area Network,' making Google's private cloud network globally accessible. The recent inclusion of Gemini 2.5 and Flash into Alphabet's Vertex AI Platform is expected to transform enterprise-AI sophistication. Amazon's cloud services are deployed via AWS. Amazon's AI applications like Alexa+ and its latest AI chip, Trainium 2, are delivering massive improvements in performance and efficiency. Amazon's Bedrock recently integrated Anthropic's Claude 3.7 Sonnet and Meta's Llama 4 model family to build high-quality Generative AI applications. The company recently launched a preview version of Amazon Nova Act, designed to help developers break down complex tasks and perform commands in web browsers. ORCL's shares have appreciated 27.7% in the year-to-date period, outperforming both the Zacks Computer and Technology sector's return of 5.5% and the Zacks Computer-Software industry's appreciation of 14.6%. Oracle's shares have also outperformed Alphabet and Amazon in the year-to-date period. While shares of GOOGL dropped 8.4%, AMZN plunged 1.0%. Image Source: Zacks Investment Research Oracle trades at a three-year EV/EBITDA of 26.53X, substantially above the Zacks Computer-Software industry average of 19.86X. ORCL has a Value Score of D. Image Source: Zacks Investment Research The Zacks Consensus Estimate for ORCL's fiscal 2026 revenues is currently pegged at $66.63 billion, indicating 16.08% year-over-year growth. The consensus mark for 2026 earnings is pegged at $6.71 per share, up 1.05% over the past 30 days. This indicates an 11.28% increase from the figure reported in the year-ago quarter. Oracle Corporation price-consensus-chart | Oracle Corporation Quote ORCL currently carries a Zacks Rank #3 (Hold). You can see the complete list of today's Zacks #1 Rank (Strong Buy) stocks here. Want the latest recommendations from Zacks Investment Research? Today, you can download 7 Best Stocks for the Next 30 Days. Click to get this free report Inc. (AMZN) : Free Stock Analysis Report Oracle Corporation (ORCL) : Free Stock Analysis Report Alphabet Inc. (GOOGL) : Free Stock Analysis Report This article originally published on Zacks Investment Research ( Zacks Investment Research Sign in to access your portfolio
Yahoo
17 hours ago
- Business
- Yahoo
Amazon takes a big hit in the AI talent wars
Amazon takes a big hit in the AI talent wars originally appeared on TheStreet. Another day and another spicy development in the AI talent wars! Meta Platform's () Scale AI deal kicked the AI talent war into overdrive, and other tech giants are following suit. 💵💰💰💵 It's an AI hiring spree, and some of the finest tech talent could rake in record-breaking AI research salaries. Lately, if a tech giant misses the boat, that's essentially letting Mark Zuckerberg's Facebook fortune write checks to poach their best generative AI talent. Needless to say, the scramble is only getting more ruthless. Over the past few months, we've seen the hunt for the top AI researchers and engineers has turned into a full-blown AI talent tech has been itching for what seems like a small pool of people who can push generative models, large language systems, and next-gen AI forward. Breakthrough AI, including lifelike chatbots like ChatGPT, Gemini, and Grok, has made it crucial to lock down top talent fast. Meta, in particular, has been especially aggressive in chasing top AI talent. Just this month, it snagged three senior researchers from OpenAI's Zurich lab, including Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai. This comes just weeks after CEO Sam Altman blasted Meta for offering nine-figure offers to poach his team. Meanwhile, Meta locked in a massive deal with Scale AI, dropping $14.8 billion for a 49% nonvoting stake. More importantly, that deal gives Meta direct access to Scale's 28-year-old CEO, Alexandr Wang, to help turn its artificial general intelligence dream into a reality. The massive Scale AI acquisition underscored the importance of valuable, top-notch, human-labeled data. Meta's big bet highlights that owning the data pipeline matters; it could potentially gain a decisive edge in building those flashy generative AI models. Meta's moves have already shaken up the whole tech space. Google's looking to cut ties with Scale to keep Meta out of its training secrets, and Microsoft's bringing in its data labeling in-house. More Tech Stock News: Circle's stock price surges after stunning CEO comment Veteran analyst drops bold new call on Nvidia stock Analyst reboots AMD stock price target on chip update Even startups like Anthropic and xAI are looking to scoop up the best talent in competing against the big guys. According to a Reuters report, Vasi Philomin, a vice president of generative AI at Amazon () , has left the company. Philomin told Reuters he's headed to another company without any was one of the major forces behind Amazon Titan and Bedrock, spending eight years shaping strategy. These pillars have become critical in Amazon Web Services' push to make AI plug-and-play for any customer. Moreover, he also helped the tech giant launch Nova for multimodal tasks and Sonic for lifelike speech, keeping pace with ChatGPT, Gemini, and Grok. He also helped plug Anthropic's Claude (from Amazon's $8 billion Anthropic investment) into Alexa, showing off Amazon's powerful playbook. Philomin's exit comes at a time when the top AI talent turns into tech's priciest prize. The top tech leaders are reportedly using sports-style scouting to find hidden talent and dropping $100 million signing bonuses to attract talent. For Amazon, losing one of its key AI architects is a big wake-up call. Despite pouring billions into big-name deals and research efforts, Amazon is at risk of its rivals snapping up its generative AI talent. Though Rajesh Sheth (formerly of Elastic Block Store) is already handling the bulk of Philomin's work, the need for a deep bench of AI leaders ready to shape models, products, and strategy is as imperative as ever. Big picture? Philomin's move is just another sign that the top minds in the AI space hold all the cards right takes a big hit in the AI talent wars first appeared on TheStreet on Jun 27, 2025 This story was originally reported by TheStreet on Jun 27, 2025, where it first appeared.
