Here's the one thing you should never outsource to an AI model - Venture Beat
In the rush to embrace new technologies, there's a loomingriskof losing the very essence of what makes truly breakthrough innovations — and, worse …
venturebeat.com 134
Covers: Hallucinations & FabricationChatbots & LLMsCybersecurityGovernance & Risk ManagementAI Safety & Existential RiskAI Agents & AutomationRegulation & PolicyMarkets & InvestmentJobs & the EconomyBanking & Finance
The AI gold rush: Whyrisksand rewards remain a balancing act | Venture Beat
Enterprises may be takingrisksby deploying AI solutions without fully considering the ethical, governance and security implications …
AWS Bedrock upgrades to add model teaching,hallucinationdetector - Venture Beat
Amazon Bedrock Model Distillation will let users use a larger AI model to train a smaller model and offer enterprises access to a model they feel …
The graph database arms race: How Microsoft and rivals are revolutionizing cybersecurity
Threat Detection: Identify high-riskattack paths, reducing false positives. Knowledge Discovery: Use graph AI for insights into interconnectedrisks …
Unlocking generative AI'strue value: a guide to measuring ROI | Venture Beat
Riskassessment and mitigation. Every gen AI project carriesrisks, from technical challenges to ethical considerations. Identifying theserisks …
This startup's AI platform could replace 90% of your accounting tasks—here's how
By maintaining human oversight, Puzzle mitigates the risk of AI errors, or “hallucinations,” as Orloff called them. “AI canhallucinate, but humans …
Service Now rolls out enterprise AI governance capabilities to accelerate production deployment
With generative AI in particular, there has been some hesitation for enterprises about safety and concerns about potentialhallucinations. Today, the …
Hallucinations & FabricationGovernance & Risk ManagementAI Safety & Existential RiskChatbots & LLMs
Securing the AI frontier: Protecting enterprise systems againstAI-driventhreats- Venture Beat
It's the weaponized AI attacks targeting identities, unseen and often the most costly to recover from that most threaten enterprises …
Identity management in 2025: 4 ways security teams can address gaps andrisks
Venture Beat“What we have to do is make sure that we use AI natively for defenses because you cannot go out and fight those AI weaponization attacks from …
Google Cloud's security chief warns: Cyber defenses must evolve to counter AI abuses
Otherrisksinclude the leaking of sensitive data via a model's output, tainting of models that can allow for prompt manipulation and biases as a …
Patronus AI launches world's first self-serve API to stop AI hallucinations| Venture Beat
A financial AI invents market data. A healthcare bot provides dangerous medical advice. These AI hallucination s, once dismissed as amusing quirks, …
Study finds LLMs can identify their own mistakes - Venture Beat
Most previous research onhallucinationshas focused on analyzing the external behavior of LLMs and examining how users perceive these errors. However …
LinkedIn founder Reid Hoffman unveils 'super agency' vision at TEDAIconference, takes ...
LinkedIn co-founder Reid Hoffman introduces 'super agency' framework for AI development, dismissing near-term electionriskswhile advocating for …
The AI edge in cybersecurity: Predictive tools aim to slash response times | Venture Beat
What role does AI play in cybersecurity? · Ransomware attacks: These hacks usually take 4 hours, but advanced persistentthreatscan take over a …
Anthropic just made it harder for AI to go rogue with its updated safety policy | Venture Beat
Anthropic updates its Responsible Scaling Policy, introducing new safety standards and AI capability thresholds to managerisksfrom powerful AI …
Prime rethinks enterprise security by design with AI systemriskanalysis and suggested actions
Prime's primary differentiator, according to Nov, is its ability to provide not onlyriskidentification but also actionable recommendations …
AI pioneer Geoffrey Hinton, who warned of X-risk, wins Nobel Prize in Physics - Venture Beat
...dangersposed by uncontrolled AI development. Hinton has warned that rapid advancements in AI could lead to unintended and harmful consequences …
How open-source LLMs enable security teams to stay ahead of evolvingthreats
From new generative AI applications to advanced security tools, these models are proving the foundation of the future of genAI-based cybersecurity …
Vera AI launches 'AIGateway' to help companies safely scale AI without therisks
Vera AI launches itsAIGateway platform, offering businesses customizable guardrails and model routing to accelerate safe and responsible AI …
Onboarding the AI workforce: How digital agents will redefine work itself | Venture Beat
To fully realize AI's potential while mitigating its risks, businesses ... Companies that delay using AI riskfalling behind in both innovation and …
Why AI is a know-it-all know nothing - Venture Beat
Right now AI systems regularly mess up, or “hallucinate” in ways that keep the mask slipping. But as the illusion of justification becomes more …
Why Microsoft's Security Initiative and Apple's Cloud Privacy Matter | Venture Beat
Apple PCC: Designed for secure AI cloud processing, it aims to prevent data leakage, insiderthreats, and targeted attacks, with robust measures to …
Building and securing a governed AI infrastructure for the future - Venture Beat
“Packaged or integrated AI tools do come withrisks, including biases in the AI models, data privacy issues and the potential for misuse.” A robust AI …
Microsoft unveils 'trustworthy AI' features to fixhallucinationsand boost privacy
Combating AI hallucination s... This tool aims to address the problem of AI hallucination s— instances where AI models generate false or misleading …
Adversarial attacks on AI models are rising: what should you do now? - Venture Beat
...AI and ML to detect networkthreatsand protect network ...AI, are automating API discovery and ending malicious bots to mitigate theserisks …
Uniphore unveils X-Stream, a unified knowledge offering to build RAG apps 8x faster | Venture Beat
With the rise of generative AI, the idea of RAG, where AI uses ...hallucination-free responses and ultimately drive efficiencies across different …
Data Gemma: Google's open AI models mitigatehallucinationon statistical queries
When tested, Google Data Gemma, tapping two distinct approaches, was able to answer several queries with high factual accuracy …
Yann LeCun slams California's SB 1047 as Geoffrey Hinton backs new regulations
The letter, submitted to Governor Newsom on September 9th, urged him to sign SB 1047 into law, citing potential “severerisks” posed by powerful AI …
Datricks gets $15M from SAP and others forAI-poweredriskand compliance platform
Datricks provides continuous, real-time monitoring of 100% of an organization's financial data, ensuring greater accuracy …
JFrog announces new integrations with Github Copilot, Nvidia Microservices and unified ops platform
Venture Beat Additionally, JFrog has introduced support for GitHub Copilot, a tool that uses AI to offer contextual coding assistance, boosting developer …