Content Fans

All News

2680 articles • Page 81 of 179

Palo Alto Networks: AI Speeds Cyberattacks 4X, Firms Have 5 Months
AI News & Trends

Palo Alto Networks: AI Speeds Cyberattacks 4X, Firms Have 5 Months

Palo Alto Networks warns that cyberattacks using AI are becoming much faster and may become normal soon. Their research suggests attackers can now steal data four times faster than a year ago by using AI tools. Experts say companies might have only three to five months before these new attack methods are common. The evidence suggests large-scale, fully automatic attacks are still being tested, but the time for organizations to prepare may be running out. Security teams that work together and focus on identity protection may be better able to stop these new types of attacks.

OpenAI unveils disaggregated Realtime-2 voice models for enterprises
AI News & Trends

OpenAI unveils disaggregated Realtime-2 voice models for enterprises

OpenAI has introduced new voice models called Realtime-2, Realtime-Translate, and Realtime-Whisper, which may help enterprises by splitting tasks like reasoning, translation, and transcription into separate parts. This separation might let companies control costs and speed, and different industries are already testing the models for things like customer calls, global support, and medical notes. OpenAI's new Deployment Company may help customers use these voice tools in their daily work. The 128K token capacity of Realtime-2 is likely more than most calls need, and developers are advised to watch for higher costs and delays if they use too many tokens. These new models suggest companies can build flexible voice systems without having to rebuild everything from scratch.

YouTube expands AI deepfake detection to all adult users
AI News & Trends

YouTube expands AI deepfake detection to all adult users

YouTube is making its AI deepfake detection tool available to all adult users, letting people aged 18 and over check if their face is being misused in AI-generated videos. Users may submit a selfie, and YouTube compares it to new uploads; if a likely match appears, users get an alert and may request removal if the use seems unauthorized. The number of takedown requests has reportedly stayed very small, which might mean few people know about the tool or it is working well so far. This change suggests YouTube is shifting toward letting regular users, not just famous people, help find possible impersonations. There may still be challenges, such as false positives and the need for more technology to catch deepfakes early.

AI accelerates cyberattacks, forcing security teams to adapt now
AI News & Trends

AI accelerates cyberattacks, forcing security teams to adapt now

AI may be speeding up both cyberattacks and defenses, as security firms report that attackers use AI to find weaknesses and spread faster. Defensive AI models also seem to find bugs more quickly than people, but human checking is still needed to avoid mistakes. The types of jobs in cybersecurity are changing as AI takes over repetitive tasks, and experts say new skills are needed. There are also worries about who is responsible if fully automated cyber tools are used, and companies might need better ways to manage risks as threats appear and change more quickly.

OpenAI Launches $4B Enterprise AI Deployment Unit, Acquires Tomoro
AI News & Trends

OpenAI Launches $4B Enterprise AI Deployment Unit, Acquires Tomoro

OpenAI has started a new $4 billion unit to help companies use AI in their most important work. The unit is supported by several investors and consulting firms, and OpenAI plans to add about 150 engineers by buying Tomoro, if regulators approve. The goal appears to be helping businesses find the best uses for AI and connect it to their own systems. Some experts think this move may increase competition and dependence on model providers like OpenAI. Success might look like more companies using AI in banking, retail, and telecom, but there could be challenges with integration and scaling.

Anthropic Urges Chip Export Controls to Maintain US AI Lead by 2028
AI News & Trends

Anthropic Urges Chip Export Controls to Maintain US AI Lead by 2028

Anthropic's policy paper suggests the U.S. should tighten export controls on advanced chips to help keep its lead in artificial intelligence by 2028. The company warns that foreign actors may be using fake accounts to copy U.S. AI models, which weakens current chip export rules. U.S. officials appear to be responding with stricter export rules and new laws that might require more checks on where chips go. Anthropic also recommends more ways to block large-scale copying of AI models and better tracking of exported chips. Experts say the exact rules and their effects are still being discussed and may change.

OpenAI Launches Personal Finance Tools for ChatGPT Pro Users
AI News & Trends

OpenAI Launches Personal Finance Tools for ChatGPT Pro Users

OpenAI has launched new personal finance tools for US ChatGPT Pro users, allowing them to connect their bank accounts through Plaid and see live financial data in their AI chats. This feature may help users get more accurate answers about their spending, budgeting, and subscriptions. OpenAI says the tool only has read-only access, and users can disconnect at any time, with data usually deleted within 30 days. There are concerns about privacy and how long data is kept, and experts suggest these issues might affect how many people use the new tools. Whether people will trust AI with their money management may depend on how well OpenAI handles privacy and accuracy questions.

EY withdraws study after GPTZero finds AI hallucinations, fake citations
Business & Ethical AI

EY withdraws study after GPTZero finds AI hallucinations, fake citations

Ernst & Young (EY) withdrew a study about loyalty rewards after reviewers, including the AI tool GPTZero, found fake citations and possibly made-up data. Some claims in the report, like the size of the loyalty-points market and fraud rates, could not be traced to real sources or seemed inconsistent. EY said it is investigating how this happened and stressed its commitment to using AI responsibly. Experts say that errors like these may risk spreading false information, especially when trusted firms are involved. The incident suggests that companies are starting to add more checks to AI-generated work, like human reviews and source tracking.

Swift, McConaughey use trademarks to combat AI deepfakes
Personal Influence & Brand

Swift, McConaughey use trademarks to combat AI deepfakes

Matthew McConaughey and Taylor Swift are using trademark law to try to protect their voices, photos, and catchphrases from being copied by AI deepfakes. Their legal filings may help stop companies from using their voices or images in ads without permission. Experts suggest these trademarks might work for celebrities with well-known brands, but it is uncertain if courts will agree that deepfakes always cause trademark harm. The law may not cover all types of deepfakes, especially those that are not used to sell products. There are also still questions about how the government will handle these new kinds of trademarks.

FINRA 2026 Mandates AI Agent Traceability for Financial Firms
Business & Ethical AI

FINRA 2026 Mandates AI Agent Traceability for Financial Firms

FINRA's 2026 rules say that financial firms using AI agents must be able to trace and prove what those agents do, not just promise they are following the rules. Firms may need to keep detailed logs of each agent's actions, especially for high-risk uses like credit or fraud, and have humans approve sensitive decisions. There might be new requirements for how data is accessed and stored, with clear limits and records about data use. Regular monitoring and checks for errors or unexpected behavior are expected once agents are active. This suggests future audits may require live proof that the firm's controls and rules actually worked each time an agent made a decision.

Harvard study finds OpenAI's o1-preview outperforms doctors in ER diagnoses
AI News & Trends

Harvard study finds OpenAI's o1-preview outperforms doctors in ER diagnoses

A Harvard and Beth Israel study suggests OpenAI's o1-preview language model may list correct or very close emergency-room diagnoses more often than doctors, especially when information is limited. The model seems to work best in situations with the most uncertainty, but experts warn that being good at tests does not mean it is ready for real patient care. Researchers say more trials and stricter safety rules are needed before using it in hospitals. Some studies also show risks if these systems are used without enough oversight. Future research will need to see how well this tool actually helps patients and doctors in real situations.

UK AI Institute Says Anthropic's Mythos Finds Critical Software Flaws
AI News & Trends

UK AI Institute Says Anthropic's Mythos Finds Critical Software Flaws

The UK AI Security Institute reports that Anthropic's Mythos AI model shows a much greater ability to find and exploit new software flaws than past models. The model may uncover weaknesses much faster than human experts, and could lead to more attacks against unpatched systems. Officials warn that Mythos appears to be improving quickly, possibly doubling its capability every four months. Experts suggest that security teams may need to adopt faster patching and better monitoring to keep up. These findings may mean bigger budgets for cybersecurity and more focus on AI oversight, but predictions depend on early test results and could change.

AI attack tools surge: 70 open-source options now available
AI News & Trends

AI attack tools surge: 70 open-source options now available

AI-powered attack tools are growing fast, with about 70 open-source options now available. Security experts warn that attackers may be using AI to find and exploit vulnerabilities more quickly, while defenders are trying to keep up. Recent reports suggest that AI agents can be both targets and tools for hackers, and that some systems may be easy to take over if not properly secured. Defensive AI can find some bugs faster than people, but still misses tricky problems, so human experts are still needed. Experts suggest that organizations should be careful with AI tools, use strong authentication, and watch for suspicious activity, as the risks may continue to increase.

UK AI Institute: AI cyber capabilities double every 4.7 months
AI News & Trends

UK AI Institute: AI cyber capabilities double every 4.7 months

The UK AI Safety Institute reports that AI models' cyber capabilities may be doubling every 4.7 months since late 2024. Their tests suggest these models can now solve more complex cyber tasks much faster, though real-world conditions might be harder. This fast progress may mean defenders need to update their security much more often, not just once a year. The institute notes that no single test proves attackers' success, but the speed of improvement suggests early and strong defenses are needed. Future experiments with tougher tests and active defenders might change how security teams prepare.

Amazon integrates Alexa for Shopping into search, replaces Rufus
AI News & Trends

Amazon integrates Alexa for Shopping into search, replaces Rufus

Amazon is replacing its Rufus assistant with Alexa for Shopping, which may make searching and shopping on Amazon more conversational. Alexa for Shopping appears to use past data from Rufus and can remember previous searches, compare products, and track prices for users. Amazon suggests this change might help shoppers by giving more personal recommendations and making it easier to buy things across devices. The move could also change how ads are shown, as Alexa for Shopping can answer natural questions and possibly keep users more engaged. It is rolling out in the U.S. soon, but there are no public numbers yet on how much this helps people buy more.