Eigen Technologies’ Post

GovGPT is an initiative to bring ChatGPT internal, more confidential, cost efficient and ingest internal data. This week's AI RiskStorming session had the team sit togeter to identify risks, mitigation strategies and... potentially larger problems with the initiative. Read the full (anonimised) report at the bottom of the blogpost in comments.

Headlines like this will become more common as AI usage scales among consumers, but it's not just the government that cares. You can't get approved by your bank if you aren't compliant with safety laws. Having started in offering infrastructure for consumer AI chatbots, we at Gabber solved the problem Senate Bill No. 243 regulates. The system we came up with keeps even the most "exciting" AI chatbot use cases compliant with banks (and now governments). And it's not as simple as moderating certain words or topics. There are four primary factors that we had to get right. - Specific word restrictions - Topical Restrictions - Prompt level guardrails - Zero-added latency These were all table stakes to creating a system that's compliant with banks and government, but still fast and functional for end users. If you want to learn more, check out this short blog post I wrote: https://lnkd.in/gGXX5ymj #AI #moderation #AIsafety #aicompliance

How seriously are you thinking about what you're pasting into ChatGPT? You could be training it to help your competition. Your AI strategy is only as strong as your data protections. Every time you skip a trust layer, feed unmasked data into public tools, or rely on bolted-on AI, your IP walks out the door. And once it’s out, it’s not just gone. It might be training a model to answer your competitor’s next product question. The future of GenAI isn’t just about speed. It’s about control, context, and compliance. Time to stop trading privacy for productivity. Learn how the Salesforce Trust Layer provides best-in-class AI security >> https://lnkd.in/essAhN4X Explore Propel One here >> https://lnkd.in/ewzycJ7K

💥 AI is rewriting the rulebook for financial risk — and most institutions are still playing by the old one. AI Introduces New Risk Architecture: The pervasive adoption of AI in financial services introduces a complex new risk architecture that traditional frameworks struggle to address. Key challenges include algorithmic biases, opaque decision-making processes, systemic reliance on third-party models, and vulnerabilities that can be exploited by adversaries. 🚨 The financial sector is experiencing a nearly threefold increase in cyberattacks, with 64% of executives believing AI exacerbates cybersecurity challenges. 👁️ Global regulators, including the EU with its AI Act, are rapidly establishing frameworks to mandate explainability, bias testing, and human oversight for AI systems. 🛡️ Data privacy and security remain a significant concern, with 72% of financial services executives citing it as a major challenge, leading some major banks to restrict employee use of external AI tools like ChatGPT. Full article in comments 👇

Financial advisors lost clients to free AI advice. The bot doesn't understand markets. It just predicts the next likely word about money. This highlights a bigger problem. People share sensitive information with AI without understanding the risks. Privacy experts warn against discussing 8 critical topics with ChatGPT: • Personal security details (passwords, banking info) • Medical diagnoses or health decisions • Emergency situations requiring immediate action • Real-time news or breaking events • Confidential business information • Legal advice for serious matters • Personal secrets or private conversations • Financial planning decisions Here's the reality. AI chatbots collect your data. They use it to improve their models. Your conversations aren't truly private. Recent vulnerabilities in AI browsers show how session hijacking can expose confidential chats. Your sensitive information could be accessed by others. Treat AI conversations like public discussions. Never share anything you wouldn't want others to know. AI lacks contextual understanding. It can't provide legitimate medical diagnoses. It doesn't have access to current market data. For sensitive decisions, consult qualified professionals. Use AI as a starting point, not the final answer. The conversational interface makes AI feel human. But underneath, it's just an algorithm with significant limitations. What's one thing you've learned to never ask AI? Share your experience below. #AIPrivacy #DataSecurity #DigitalSafety 𝐒𝐨𝐮𝐫𝐜𝐞: https://lnkd.in/gfWxuZKJ

The AI revolution is here, but so are the compliance challenges. 🚨 If you're deploying AI in your organization, you need to watch our new video. We're dissecting the compliance frameworks of major AI players. From Microsoft Copilot to ChatGPT, Claude, Abacus.AI, and Perplexity. Learn how to mitigate risks, avoid data breaches from "shadow IT," and ensure your AI adoption meets regulatory standards like SOC 2, HIPAA, and SEC. Protect your business and leverage AI responsibly. Full breakdown here: https://lnkd.in/g7vdfZbC #AIStrategy #RiskManagement #ComplianceMatters #EnterpriseTechnology #DigitalTransformation #FutureofWork #TechInsights

The AI revolution is here, but so are the compliance challenges. 🚨 If you're deploying AI in your organization, you need to watch our new video. We're dissecting the compliance frameworks of major AI players. From Microsoft Copilot to ChatGPT, Claude, Abacus.AI, and Perplexity. Learn how to mitigate risks, avoid data breaches from "shadow IT," and ensure your AI adoption meets regulatory standards like SOC 2, HIPAA, and SEC. Protect your business and leverage AI responsibly. Full breakdown here: https://lnkd.in/g7vdfZbC #AIStrategy #RiskManagement #ComplianceMatters #EnterpriseTechnology #DigitalTransformation #FutureofWork #TechInsights

🔲 ShadowLeak shows how fast the rules of security are shifting. Radware researchers demonstrated that a simple hidden prompt in an email could trick ChatGPT into exfiltrating inbox data, invisibly👻 . AI-driven integrations open new blind spots, and application security needs to evolve to keep pace. What do you see as the biggest risk with AI agents in security? Source: Dark Reading Article link in comments below⬇️ #appsec #ai #shadowleak #news (via AppSec Village)

Shadow AI - the growing challenge that's keeping CISOs up at night. Employees across your organisation are using ChatGPT, Claude, and dozens of other AI tools to boost productivity, often feeding them sensitive company data, proprietary code, and confidential information without realising the risks. Once that data leaves your environment, you've lost control over it, creating massive exposure for data leakage, compliance violations, and intellectual property theft. How are you dealing with Shadow AI in your organisation?

ChatGPT can't legally give financial advice. Young investors ask it anyway. The gap between what's allowed and what's happening is massive. The numbers tell the story. 80% of young investors find AI helpful for finance decisions. But half admit it led to poor choices. That's the danger of using general AI tools for regulated activities. WNSTN AI built something different. Three specialized agents working together: •Engagement Agent: Portfolio reviews and education •Compliance Agent: Regulatory oversight •Intelligence Agent: Broker insights This isn't just another chatbot. Every interaction is recordable and auditable. For brokers, this solves a real problem. Young investors expect personalized, real-time information. WNSTN delivers that while keeping compliance intact. The business intelligence piece is smart too. Track what investors ask about. See trending topics. Tailor your offerings accordingly. As AI regulations tighten globally, WNSTN shows how innovation and compliance can work together. Investors get reliable information. Brokers deploy AI features rapidly. Regulators see responsible implementation. Win-win-win. The future of financial AI isn't about replacing human judgment. It's about augmenting it with proper guardrails. How do you see AI changing financial advice in your industry? #FinTech #AIInFinance #ComplianceFirst 𝐒𝐨𝐮𝐫𝐜𝐞: https://lnkd.in/dmDmBN5s