Latest news with #TechCrunch
Indian Express
an hour ago
- Science
- Indian Express
Tracking AI models' ‘thoughts' could reveal how they make decisions, researchers say
A broad coalition drawn from the ranks of multiple AI companies, universities, and non-profit organisations have called for deeper scrutiny of AI reasoning models, particularly their 'thoughts' or reasoning traces. In a new position paper published on Tuesday, July 15, the authors said that monitoring the chains-of-thought (CoT) by AI reasoning models could be pivotal to keeping AI agents in check. Reasoning models such as OpenAI's o3 differ from large language models (LLMs) such as GPT-4o as the former is said to follow an externalised process where they work out the problem step-by-step before generating an answer, according to a report by TechCrunch. Reasoning models can be used to perform tasks such as solving complex math and science problems. They also serve as the underlying technology for AI agents capable of autonomously accessing the internet, visiting websites, making hotel reservations, etc, on behalf of users. This push to advance AI safety research could help shed light on how AI reasoning models work, an area that remains poorly understood despite these models reportedly improving the overall performance of AI on benchmarks. 'CoT monitoring presents a valuable addition to safety measures for frontier AI, offering a rare glimpse into how AI agents make decisions,' the paper reads. 'Yet, there is no guarantee that the current degree of visibility will persist. We encourage the research community and frontier AI developers to make the best use of CoT monitorability and study how it can be preserved,' it adds. The paper calls on leading AI model developers to determine whether CoT reasoning is 'monitorable' and to track its monitorability. It urges deeper research on the factors that could shed more light on how these AI models arrive at answers. AI developers should also look into whether CoT reasoning can be used as a safeguard to prevent AI-related harms, as per the document. But, the paper carries a cautionary note as well. It suggests that any interventions should not make the AI reasoning models less transparent or reliable. In September last year, OpenAI released a preview of its first-ever AI reasoning model called o1. This launch prompted other companies to release competing models with similar capabilities such as Gemini 2.0, Claude 3.7 Sonnet, and xAI's Grok 3, among others. Anthropic researchers have been studying AI reasoning models, with a recent academic study suggesting that AI models can fake CoT reasoning. Another research paper from OpenAI found that CoT monitoring could enable better alignment of AI models with human behaviour and values.
The Verge
an hour ago
- The Verge
Google's Discover page now summarizes news with AI
Google is rolling out AI-generated summaries directly within Discover, its personalized news feed nestled within the Google Search app. TechCrunch reports that some users in the US are seeing Discover cards on iOS and Android that provide the new summaries in place of a headline and logo from a single publisher, similar to how AI Overviews appear at the top of results in Search. It acts to further obscure news sources at a time when search traffic to publishers is disappearing. The AI summaries in Discover display overlapping icons in the top-left corner, indicating the number of cited stories. Tapping the icons reveals a 'More' sheet with all the contributing articles. Users get a quick three-line preview before needing to tap 'See more' to expand the summary. A prominent warning is displayed underneath the text disclosing that the summaries are generated with AI, 'which can make mistakes.' This feature appears to have been in development for a while. Google bug hunter AndellDam found that trending cards with AI Overview-like summaries were being tested on Google Discover in June. A Google spokesperson confirmed to TechCrunch that this is not a test, but a US launch on iOS and Android. Google told the publication that the feature will focus on trending lifestyle topics like sports and entertainment, and aims to make it easier for people to decide what pages they want to visit. This is the latest of several AI features Google has been toying with that discourage users from clicking on web links, following AI Overviews, Audio Overviews, and AI Mode being tested in Search. This is cause for concern across the publishing industry, which has seen a dramatic fall in website traffic and referrals now that AI is pulling information out of articles and negating the need for readers to click through to original sources. That loss of traffic and revenue from Google Zero has already contributed to killing sites like AnandTech, Giant Freakin Robot, Laptop Mag, and Buzzfeed News — reducing the pool that Google can feed to its AI.
Hans India
2 hours ago
- Hans India
Meta Fixes AI Privacy Bug That Exposed User Chats, Awards ₹8.5 Lakh to Ethical Hacker
Meta has resolved a critical privacy flaw in its AI chatbot platform that could have exposed users' private conversations to malicious actors. The vulnerability, flagged late last year, was responsibly disclosed by ethical hacker Sandeep Hodkasia, who was awarded a bug bounty of $10,000 (roughly ₹8.5 lakh) for his discovery. According to a report by TechCrunch, Hodkasia—founder of the cybersecurity firm AppSecure—reported the issue to Meta on December 26, 2024. The flaw, linked to the prompt editing feature in Meta's AI assistant, had the potential to allow unauthorized access to personal prompts and responses from other users. Meta users interacting with the AI platform can edit or regenerate prompts. These prompts, along with AI-generated replies, are each assigned a unique identification number (ID) by Meta's backend system. Hodkasia found that these IDs, which were visible through browser developer tools, followed a predictable pattern and were vulnerable to manipulation. 'I was able to view prompts and responses of other users by manually changing the ID in the browser's network activity panel,' Hodkasia explained. The major issue, he pointed out, was that Meta's system didn't verify whether the requester of a particular prompt actually owned it. That meant someone with modest technical knowledge could write a script to cycle through IDs, collecting sensitive user data at scale. The ease with which this vulnerability could be exploited made it particularly dangerous. Since the system lacked user-specific access checks, it effectively opened a backdoor to private AI conversations. Thankfully, Hodkasia chose to report the issue rather than exploit it. Meta confirmed it patched the flaw on January 24, 2025, following an internal review. The company also stated that there was no evidence suggesting the vulnerability had been exploited before Hodkasia's report. While the fix has been deployed, the incident has renewed concerns about data privacy in AI platforms. As tech giants race to roll out AI-powered products to stay ahead of the competition, lapses like this highlight the growing importance of robust security protocols. Meta launched its AI assistant and a standalone app earlier this year to compete with platforms like ChatGPT. However, its rollout has not been without issues. In recent months, some users reported that their supposedly private conversations were visible in the platform's public Discovery feed. Although Meta maintains that chats are private by default and only become public when explicitly shared, users argue that the app's interface and settings are confusing. Many claimed they were unaware that their personal inputs, including photos or prompts, might become publicly accessible. As AI tools become more integrated into daily life, incidents like this serve as a stark reminder of the need for transparency, user control, and stringent privacy protections. Meta's swift response and bug bounty program underscore the critical role of ethical hackers in maintaining digital safety.
Mint
4 hours ago
- Mint
Google is rolling out AI summaries in Discover feed: Know what it means for users
Google is reportedly rolling out a new AI-powered feature to Discover, its news feed on Google's search app. Some users have started to see AI summaries of top news articles on the Discover feed instead of headlines from publications, as per a TechCrunch report. The feature has started to roll out on the Google app for both Android and iOS devices. The AI summary of news articles is being generated from several sources, with publishers' logos displayed in the left corner of a particular news article. While this feature may roll out as an easy way user users to get the latest news at a glance, but it could put publishers in major traffic trouble. Google is said to bring AI summaries to the Discover feed for all the trending news articles. These summaries are likely formed using information from the publishers, which are also listed on the left side of the summary. Just below the AI summary, Google also warns that the content is 'generated with AI, which can make mistakes.' As of now, only a few users have access to this feature, leading users in the US to speculate as a test on Android and iOS. However, TechCrunch reported that a Google Spokesperson has confirmed that the AI summary is officially rolling out to users in the U.S. It was further highlighted that the AI summaries will be focused on lifestyle-focused articles such as sports and entertainment. This will allow users to decide which page or publication they want to visit for detailed information. Alongside AI summaries, some news articles all also being displayed in bullet points, but they are not flagged as generated by AI. However, Google might be testing new ways to display information on the Discover feed, which makes it easier for users to get the gist of information. The AI summaries in Discover are currently not available in India or other markets except the U.S. However, the tech giant may plan to roll out the feature globally later.
Business Insider
6 hours ago
- Business
- Business Insider
Experts reveal why media mentions are the new SEO currency for growing brands
Experts reveal why media mentions are the new SEO currency for growing brands Experts say backlinks alone no longer cut it — and great content also alone isn't what gets you ranked. What is it then? Being talked about. Quietly, media mentions have become one of the strongest trust signals in Google's eyes. Here's why they're now the real currency of modern SEO. Google Ranks Brands, Not Just Websites For years, SEO was about keywords, metadata, and backlink profiles. But Google's algorithm is no longer just crawling for technical signals… It's evaluating identity. Whether it's a quote in TechCrunch, a founder profile in a niche blog, or your product name in an industry roundup, these mentions help Google connect the dots. It helps map entities in Google's Knowledge Graph. Which is how your brand becomes recognizable in the search ecosystem. 'The algorithm doesn't just care about backlinks anymore,' explains Htet Aung Shine, Co-Founder of NextClinic. 'It's about authority, and authority comes from consistent signals across multiple trusted platforms; media mentions being one of the strongest.' This kind of authority can't be faked with SEO tricks. It's earned. When trusted sources mention your brand — even casually — they're reinforcing your legitimacy. Over time, these signals create a public presence that algorithms recognize and reward. 'Think of every mention as a vote of confidence,' says Sinead Corceran, Yoga Trainer ERYT200 & Course Director at All Yoga Training. 'When Google sees others talking about you — especially in context — it assumes you're worth showing to more users.' Unlinked Mentions Are Still Powerful (and You're Probably Ignoring Them) There's a trap many SEO teams fall into: assuming that if a brand mention doesn't come with a blue, clickable backlink, it's worthless. Well, that's no longer the case. Because Google has shifted from link-counting to entity understanding. It uses mentions to confirm that your brand exists, is relevant in your niche, and is part of ongoing conversations across the web. 'We've tested pages with no backlinks, just unlinked brand citations — and they still performed well,' says Martin Seeley, CEO & Senior Sleep Expert at Mattress Next Day.' When your brand name keeps appearing in credible sources — even without a link — Google begins to connect it with specific topics, keywords, and categories. That helps you rank for more than just your own content. You start to appear in broader search contexts because you've become part of the digital narrative. 'If someone reads your name on a major media site, that alone builds familiarity,' says LJ Tabango, Founder & CEO of Leak Experts USA. 'Even without a link, they might Google you. That brand search gives Google another trust signal.' Rightfully so. Unlinked mentions influence human behavior, too. Readers come across your brand in articles or social posts, and even without a link, they search your name directly. One Mention in the Right Place Can Outperform 100 Low-Quality Backlinks 'Dozens of backlinks from average blogs might look impressive in a spreadsheet, but in reality, a single media mention from the right source can deliver 10x more impact — both for rankings and brand perception,' says Stephen Twomey, Founder of MasterMindSEO. The difference comes down to context and credibility. A mention in an outlet like Forbes, TechCrunch, or even a respected niche publication carries more weight because it's both editorial and earned. Google knows these outlets don't link to just anyone. So when they reference your brand, especially within content that matches your niche, it signals serious trust. Paula Mixides, Link Building Expert at Cake Box - Birthday Cakes, says, 'A high-authority mention in a relevant article can do more for your SEO than dozens of links from directories or generic blogs.' That ripple effect is something low-effort links simply can't replicate. That is especially true for startups and growth-stage brands. Instead of pouring effort into mass link-building, consider focusing on just one high-quality media opportunity per month. 'People underestimate how powerful one well-placed editorial mention can be,' explains Kasia Siwosz, Life Coach. 'You get trust, traffic, SEO value, and brand lift — all in one.' Media Mentions Drive Branded Search — The Most Underrated SEO Metric The most trusted brands in search aren't always the ones with the most links… they're the ones people actually look for. Branded search is where that starts. When someone sees your name in a respected article and types it into Google, they're sending a powerful signal: this brand matters. That action — even once — tells the algorithm you're worth noticing. Multiply that by dozens or hundreds of users over time, and suddenly your name carries weight in the SERPs. 'Branded search volume is one of the strongest signals that you're building a real business, not just a content farm,' says Michael Jensen, CMO at Forsikringssiden. 'Google wants to rank the brands people look for — mentions create that curiosity.' These searches also improve click-through rates. A user who recognizes your brand from media coverage is more likely to click when you appear in search. That engagement reinforces rankings and conversions. 'When users type your name directly into search, that's when you know you've made it past the algorithm,' explains Michael Song, Marketing Manager at Zhoushenghang used buses. 'Media mentions spark that interest — they create the reason to search in the first place.' Journalists Are Your New Link Builders — But They Hate Traditional Pitches The best links in SEO don't come from link builders — they come from journalists. And most brands pitch them completely wrong. Journalists aren't looking for your homepage, your About page, or your PR jargon. They're looking for real, helpful insight — fast, clear, and relevant to the story they're writing. If you give them that, they'll quote you. And if they quote you, the link will usually follow. Journalists don't care about your product. They care more about whether you have something useful to say in context of the article. That's what gets included. That shift means brands need to stop acting like advertisers and start acting like contributors. Whether it's through HARO Services, Qwoted, or direct relationships, success comes from positioning your people — not your brand — as a valuable part of the conversation. 'Reporters want smart quotes from smart people,' says Rameez Usmani, Founder & CEO of HARO Link Building Service - 'If your insight fits the piece and doesn't sound self-serving, you've earned yourself a backlink without asking.' Mentions Create a Feedback Loop for Topical Authority When your brand shows up in multiple articles around the same topic, Google starts connecting you to that theme, even if you're not ranking for those keywords yet. That's how you build topical authority without publishing dozens of blog posts. A few high-quality mentions in relevant media can position you as a go-to source in a niche, and that reputation carries over into rankings. 'You don't need 100 pages about a topic — you need 10 good mentions in the right places,' says Alex L., Founder of StudyX. These mentions act like external citations. Google sees you consistently tied to certain themes — like AI tools, digital health, or sustainable fashion — and starts surfacing your content higher when those topics are searched. Magnus Larsen, Head of Marketing at Forbrukerguiden, explains, 'Those external signals help Google trust your site before your content even catches up.' It's a loop: the more mentions you earn in a niche, the more likely you are to rank. And the more you rank, the more likely you are to be mentioned again. You're Wasting Media Mentions by Not Repurposing Them Most brands treat media mentions like a one-time win — they celebrate, then forget. But smart brands squeeze every bit of value out of each one. A quote in a top-tier blog? Turn it into a LinkedIn post. A founder mentioned in a newsletter? Add it to your homepage. A customer story featured by a third party? Use it in your sales deck. Repurposing extends the shelf life of each mention — and boosts its SEO value. 'If you're not building a content loop around every media feature, you're wasting the compound effect,' says Nick Borges, COO at Heartwood Recovery. You can also link back to the original mention to boost its authority and visibility. This creates a second layer of SEO benefit — helping that article rank while reinforcing your association with it. 'Mentions shouldn't just live in your press section,' says John Beebe, Founder of Classic Car Deals. 'They should be everywhere your audience looks — sales calls, email nurture, even onboarding.' A single media mention has more value than most blog posts — if you know how to keep it alive. Founders and Experts Get Quoted — and That's Your SEO Advantage Richard Zi, CEO at ZW Cable, says, 'Media mentions are one of the fastest ways to earn trust, not just with people, but with Google. And the easiest path to those mentions is to put a real voice front and center.' Journalists quote founders, specialists, and real humans, not generic brand blurbs. When you show up as a person with a point of view, you're more likely to be mentioned, quoted, and linked in editorial content. 'People connect with people. That's why expert-driven content and quotes outperform any company boilerplate,' says Faizan Haider, SEO Specialist & Link Builder at Qwoted Link Building.
