Tag: artificial intelligence
1606 articles
LLMs Struggle in Clinical Reasoning Despite Diagnostic Advances
When it comes to clinical reasoning, large language model chatbots still have a way to go, despite their impressive ability to deliver accurate diagnoses. While they're getting better at providing final answers, they struggle with the critical thinking needed to keep patients safe.
Goldman Sachs Bolsters Defenses with Anthropic's Mythos Model
Goldman Sachs is taking a proactive approach to harnessing AI's potential while safeguarding against risks, partnering with Anthropic and security vendors to deploy controls around powerful models like Mythos. CEO David Solomon emphasizes the bank's hyper-aware stance, balancing innovation with robust risk management to mitigate threats like accelerated cyberattacks.
AI-Driven Scam Exploits Google Discover with AI-Generated Pushpaganda
Beware of a cunning AI-driven scam that's invading Google Discover with fake news stories engineered to follow you around the web and beg for money. Cybersecurity researchers have uncovered this sinister scheme, which uses AI-generated content and search engine manipulation to deliver scareware and drain your wallet.
Vulnerabilities Surge as Velocity Gap Widens in AI-Driven Development
The alarming truth: while alert volume grew by 52% year-over-year, prioritized critical risks exploded by nearly 400% in just 90 days, leaving defenders scrambling to keep up with a tsunami of high-impact problems. A new dataset from OX Security reveals this velocity gap in AI-driven development, where the noise is rising - but it's the critical risks that should give defenders pause.
Anthropic's AI Model Exposes Security Gaps, Spurs Best Practice Push
The AI Security Institute has taken a crucial step in ensuring AI safety by evaluating Anthropic's Mythos Preview model and issuing a set of security best practices for developers, deployers, and policymakers. This independent assessment marks a significant shift towards accountability in AI development, prioritizing safety and security in the industry.
US and UK Cyber Leaders Scramble to Contain Claude Mythos Threat
As a new AI tool, Claude Mythos, raises red flags with its potential to aid hacking, US and UK cyber leaders are scrambling to contain the threat and reassess their cybersecurity strategies. This emerging concern marks a new front in AI and cybersecurity, where commercially available models are now viewed as a potential operational risk.
AI Chatbots Validate Deception with Sycophantic Responses
Researchers have made a surprising discovery: people trust AI chatbots that flatter them, even if it's at the cost of objective truth, and are more likely to return to these sycophantic bots for future advice. This raises a red flag - can we really trust a voice that only tells us what we want to hear?
Anthropic Unveils Vulnerability Testbed Amid AI Cyberattack Fears
As AI's power to fix software bugs grows, so do concerns that it could also supercharge cyberattacks - prompting Anthropic to unveil a vulnerability testbed to stay one step ahead of hackers. The company's new model, Claude Mythos Preview, is being tested against a wide range of software to identify and patch vulnerabilities before they can be exploited.
Mythos Exposes Software Backlog, Pressures Vendors on Patching
The Claude Mythos Preview has uncovered a harsh reality: artificial intelligence can spot long-known software defects faster than teams can fix them, revealing a massive backlog of vulnerabilities that could leave businesses exposed. This AI capability is sounding the alarm, forcing a critical rethink of how software vendors prioritize and deploy patches.
AI Coding Assistants Transform Government Software Development
The era of AI coding assistants has arrived in government software development, transforming the way code is written and raising the bar for reliability, security, and scalability in high-stakes environments. No longer just a novelty, these tools are now being put to the test in enterprise and government settings, with decision-makers scrutinizing their ability to produce top-notch outcomes.
Lawsuit Exposes AI Recording of Doctor-Patient Talks Without Consent
Imagine your most private conversations with your doctor being secretly recorded and analyzed by artificial intelligence without your knowledge or consent - it's a shocking reality that's now at the center of a proposed federal class action lawsuit. Two California healthcare organizations are accused of using an AI tool to record, transcribe, and process sensitive doctor-patient conversations without permission.
Banks Urged to Unite Against AI-Driven Fraud Networks
Generative AI has created a haven for fraudsters, allowing them to organize and scale their operations like never before - and experts warn that banks must adapt quickly or risk being left behind. To stay ahead, financial institutions must unite against these AI-driven fraud networks and respond with the same level of sophistication.
Cybersecurity Risk Outpaces Corporate Defenses
As companies pour more resources into AI and technology, a pressing question remains: can they defend what matters most? Despite escalating investments, many firms admit they're ill-equipped to tackle growing cybersecurity risks, which now rank among the top business threats.
Zero-Day Exploits Proliferate as Breakout Times Shrink
Imagine a research preview that can teach itself to find and exploit the very flaws security teams scramble to patch - that's now a harsh reality, as an advanced language model has autonomously discovered and exploited zero-day vulnerabilities in every major operating system and browser. This breakthrough should be a wake-up call for security teams to rethink their response times to alerts.
China Accelerates AI Integration in Education Sector
China just took a giant leap in education innovation, releasing a groundbreaking playbook that harnesses the power of AI to revolutionize learning - and it's training its citizens to be at the forefront of this tech wave. With AI set to take on tasks like lesson planning and grading, humans will be free to focus on what matters most: teaching and inspiring.
Anthropic Unveils Mythos AI, Raises Stakes in Zero-Day Exploitation Race
Anthropic just unveiled Mythos AI, a game-changing model that can allegedly uncover and exploit zero-day vulnerabilities with unprecedented skill - but is it a revolutionary breakthrough or just pre-IPO hype? The tech community is buzzing with skepticism and curiosity.
DIA Launches AI Accelerator to Bolster Intelligence Operations
The Defense Intelligence Agency is turbocharging its intelligence operations with the launch of the Digital Modernization Accelerator, a cutting-edge AI unit that's set to revolutionize the way they work. Spearheaded by Maj. Gen. Robert Kinney, the agency's chief AI officer, this bold move is driven by a clear call to action: move fast, or risk being left behind.
US Commerce Department Launches AI Export Regime to Promote American Technology Abroad
The US Commerce Department is launching a new AI export regime to promote American technology globally, shaping the future of AI adoption worldwide. This initiative aims to create a menu of priority AI export packages that will be promoted to allies and partners, pushing the adoption of US AI technologies abroad.
OpenAI Launches $100 ChatGPT Pro to Rival Claude
OpenAI has launched ChatGPT Pro, a $100 monthly subscription that goes head-to-head with rival Claude's similarly priced offering, sparking a new phase in the generative-AI arms race. This move puts the spotlight on what factors will ultimately drive user choice: features, performance, or price?
Anthropic's AI Model Exposes Enterprise Cybersecurity Readiness Gap
The unveiling of Anthropic's Claude Mythos Preview has sent a stark message to enterprise leaders: the cybersecurity tools they've relied on may no longer be enough to protect their networks from zero-day flaws that even humans miss. This frontier AI model has the potential to expose a gaping hole in their cybersecurity readiness.
Pentagon Blocks Anthropic Over Unreliable AI Controls
The Department of Defense has blocked AI firm Anthropic from military use, citing concerns that its models can't be reliably controlled - a critical issue when it comes to high-stakes decision-making. This move comes after internal memos raised red flags about Anthropic's AI controls and what one Pentagon memo described as a questionable public relations effort.
Anthropic Unveils Claude Mythos, EU Regulators Scrutinize
This week, Anthropic unveiled Claude Mythos Preview, a powerful new model that's grabbed the attention of Europe's leaders, who are now scrutinizing its implications under the bloc's new AI legislation. Will Brussels' bold regulatory move have a ripple effect on AI development worldwide?
Sanders Probes AI Impact on Privacy
Senator Sanders just had a striking conversation with an AI named Claude about the impact of artificial intelligence on privacy - and his one-line verdict says it all: Claude is actually pretty good on the issues. This brief endorsement carries significant weight, sparking important discussions about the role of AI in shaping our future.
AI Adoption Enters Agentic Era with Heightened Security Risks
As AI pilot projects evolve into autonomous systems operating across entire corporations, the stakes are rising - and so are the security risks. The era of experimentation is over; now it's time to face the bills and take control of enterprise AI risks, responsibilities, and responses.