Big Data & Analytics - Thinks & Links | August 5, 2023
BD&A - Thinks and Links
Big Data & Analytics - Thinks and Links | News and insights at the intersection of cybersecurity, data, and AI
Happy Weekend!
We held several events this week with Optiv clients to discuss the implications of AI for attackers and defenders. Many of our clients are already embracing the expanded focus on AI this year, particularly ChatGPT and related Large Language Models. Leaders recognize its vast potential to transform their operations, but also have a focus on realism. In 2023, if teams had all the data they needed to safely automate a function, they’ve likely already done so. The devil is in the details, and LLMs are not a fix for every security problem. However, the promises of LLMs and related tools are fueling optimism and innovation for AI, including in the cyber security industry.
The biggest commonality among clients we met with this week as the usage of LLMs as tools that can help manage large masses of information. Knowledge management is the “low hanging fruit” of the LLM revolution. The ability of AI systems to synthesize large volumes of data is a particularly attractive attribute.
Our clients are not just considering AI’s potential, they are actively exploring ways to integrate and apply it within their organizations. Examples include knowledge bases for SOCs and acceleration of coding and content writing. The focus is on establishing AI systems that operate together with humans, because everyone agrees that these automations can’t be trusted for fully autonomous control of critical functions.
Another area where clients are particularly focused is the creation of AI policies. Central to these discussions is data governance, employee training, and access control. The consensus is that the application of AI is certainly accelerated by strong data governance capabilities, but it is also tightly tied to Identity and Access Management, Compliance, Legal, and Risk.
There’s also a big difference between experiments within IT teams and moving from proof of concept to full-fledged application. Many clients expressed the challenge of explaining AI’s risks to larger audiences, where security doesn’t want to be only in the business of saying “no.” There’s widespread agreement that there aren’t yet enough standardized guidelines and tools for all of the concerns that LLMs raise, but products and services are evolving quickly to compensate.
Despite these challenges, our clients are largely optimistic. They recognize that AI is not merely about security and technology but is also an opportunity for revenue generation and growth. We all agree there will be many bumps along the way and likely some spectacular failures. But clients who are proactive and engaged in these changes appear ready to weather these upcoming shifts.
Optiv & Business Security Weekly: How To Effectively Embrace And Protect Generative Ai Tools, Models, & Data
I truly enjoyed the conversation with hosts of the Business Security Weekly podcast. We covered a lot of my favorite topics including discussing the risks and opportunities of generative AI. We all agreed that the “horse has left the stable” but many organizations are moving at different paces when it comes to adopting and securing AI. We discussed the risks associated with using AI as well as potential mitigation strategies. Regardless of where your organization is on the AI maturity spectrum, discussing AI responsibility and learning about its potential is key.
IBM: AI reduces data breach lifecycles and costs
https://securityintelligence.com/articles/ai-reduces-data-breach-lifecycles-and-costs/
AI-based cybersecurity tools reduce financial impact of data breaches, according to IBM. Benefits include:
Great to see quantitative approach to measuring the benefits for AI in cybersecurity
Elastic: Avoid Shadow AI — Embrace generative AI in the SOC
https://www.elastic.co/blog/avoid-shadow-ai-generative-ai-soc
Generative AI offers a potential solution to the cyber skill gap, but control & security of data must be maintained. Elastic AI Assistant prioritizes transparency, allowing users to control data shared with models. Solutions like Elastic Security ensure compliance & monitoring of AI usage. Keeping data local & running smaller models locally offer benefits. Organizations should embrace & empower teams to use AI while maintaining control & security.
As Artificial Intelligence Accelerates, Cybercrime Innovates
https://www.darkreading.com/risk/as-artificial-intelligence-accelerates-cybercrime-innovates
Cybercrime has become the world’s third largest economy, according to new data from Verizon. Generative AI is set to magnify the impact, with criminals using AI to create malware, deepfakes, and phishing emails. Cybersecurity companies are innovating with AI to mitigate risk; meanwhile, industry and governments are now actively engaging in the prospects of AI and regulations. Now is the opportunity to make rapid strides to improve cybersecurity.
New OpenAI-Enabled Security Product: Security Genie from Solvo
Another day, another product delivering an OpenAI-powered LLM as a part of its roadmap. From the announcement blogpost:
Human error is the top cause of cloud security breaches and Gartner predicts it’ll be responsible for over half of major cyber incidents by 2025. Solvo’s AI-powered SecurityGenie companion helps cloud security specialists gain insight, assess compliance score & remediate code, saving 5 hours/week.
Have a Great Weekend!