#AIChat is a 60 minute Twitter chat on artificial intelligence (AI).
Where my guest cuts through the AI noise.
To give you actionable advice you can use at work.
Today’s guest? Tricia Howard. Tricia is the best when it comes to #cybersecurity. She works as @Optiv. And you can read her cybersecurity articles on her website: https://triciakickssaas.com/. Or from her Twitter account (@TriciakicksSass)
The topic? AI and cybersecurity.
Here’s question Q1) How is #AI being used in #cybersecurity? #AIChat pic.twitter.com/d5ijfrpyLZ
— Nick Tang (@nickhtang) October 24, 2018
A1 – There are two sides to this, the whitehat side and the blackhat side. From the good side, there is everything from cognitive search engines, ahem, @IBMWatson, all the way to zero-day #malware protection. One of the most practical applications is the #UEBA factor. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A1 (2) Being able to recognize anomalies in user behaviour which kicks off an automated workflow.The most exciting to me personally is the automated data lake. Data is only as good as what you use it for. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A1 (3) When we talk #BigData someone has to be there to ingest and make sense of all of it to bring out real business use cases. #AI can do that, and at a fraction of the time it would take for humans to analyze. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A 1: AI and Machine Learning are being combined to provide insights that humans can’t catch in devices – anomalies, speed up detection, and other subtle, notable patterns that could indicate a breach of security. #AIChat
— US Cybersecurity Mag (@USCyberMag) October 24, 2018
A1 – suppose your system admin’s master password was #hacked. #AI will detect that the credentials were used from Pakistan – but your admin 👩💼 lives in Miami. It will trigger a big red flag and sound the alarm bells.
— Avrohom Gottheil (@avrohomg) October 24, 2018
We agree with @USCyberMag!
“As with any new technology, there will be failures”
— Proven Data (@Proven_Data) October 24, 2018
While you’re all discussing Q1 for @TriciaKicksSaaS, let’s start on Q2. What are the challenges of using #AI in #cybersecurity? #AIChat pic.twitter.com/MNXDxAoE0d
— Nick Tang (@nickhtang) October 24, 2018
A2: This is where I was talking about the two sides of #AI. With every great innovation there can be a negative application. From the “bad” point of view – there are so many “hacker toolkits” that automate attacks at a relatively low cost.
#AIChat— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A2 (2): It is much more expensive to defend against these attacks than it is to deploy them (most of the time.) There’s still a lot of negative stigma around #AI as well which can be a challenge. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A2 (3): There are still lots of places for improvement for #AI in #security. It’s not a catch-all. There needs to be specific applications for the AI to be used in order for it to be effective… at least until we get truly cognitive ‘beings.” #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A2 The challenge for using #AI in #Cybersecurity is that #technology changes rapidly, and the threat actors are extremely advanced. We have to continue to innovate to stay ahead of the game. #aichat
— Avrohom Gottheil (@avrohomg) October 24, 2018
A2: Challenges could stem from turning over all tasks to automation. #AI still requires some level of human interaction. It’s not a complete replacement, especially when critical data & PII etc are at stake in #cybersecurity. #aichat
— Adlib Software (@adlibsoftware) October 24, 2018
Yep. there always needs to be oversight. Take a light bulb for example: it’s pretty self-sustainable, does its job, etc…until it burns out and needs to be changed. We improve the life of them and optimize them, but they still have to be changed when they’re burnt out. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
As you’re talking about Q2, here’s Q3 for @TriciaKicksSaaS: Do you think #AI will replace humans in #cybersecurity? #AIChat pic.twitter.com/EkSQZsVesB
— Nick Tang (@nickhtang) October 24, 2018
A3: Aha! One of the negative stigmas! I’ve actually written about this (not specific to security) and the answer is no. They will augment humans but not replace them. We will just get more sophisticated – more analytical roles rather than basic blocking and tackling. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A3 (2): True #AI is set it and forget it – and some of this is still rudimentary. We still need people tuning the software until it truly is learning on its own. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A3 (3): Aside from that, as we’ve said here today, there still needs to be oversight and coaching to get the machine to do the intended job. We don’t want to hinder it persay, but we need to be sure we’re keeping an eye on where it’s going to stay relevant. #AIchat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
Along the line wont the capabilities/ problems ahead allow for solely #ai intervention ??? In times when #ai could write its code / modify ( still proto ) just imagine capabilities used as exploitation tool ??? #CyberSecurityAwarenessMonth #CyberSec #cyberattacks #aichat
— Jan Barbosa 🐝 (@JBarbosaPR) October 24, 2018
I wouldn’t say solely #AI, but yes machines will be battling each other. It’s happening now.The bad peeps are using a code to figure out how to get in and the humans are catching up. When the machine becomes truly cognitive there will be AI battles. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A3: The #futureofwork will likely be augmented by #AI, not obliterated because of it. Teams can focus on strategy while #AI focuses on the automated, mundane tasks. #AIchat
— Adlib Software (@adlibsoftware) October 24, 2018
#AI systems yield a large number of false positives. You eventually need the help of human experts to find evidence of security breaches and make critical decisions. Till we attain #singularity 😂
— Archonsec (@archonsec) October 24, 2018
A3) As we have talked about extensively today, #AI has a wide margin for error, and therefore will never fully replace humans. Even when AI is perfected, we simply will have a different set of responsibilities and capabilities. #AIChat
— US Cybersecurity Mag (@USCyberMag) October 24, 2018
A3. Certainly not. #AI will only add up to human resources rather than replacing them for #cybersecurity
AI will be an enabler rather than a replacement for humans when it comes to cyber security.#AIChat
— Moin Shaikh (@moingshaikh) October 24, 2018
Last 10 mins of #AIChat with @TriciaKicksSaaS Talking #AI and #cybersecurity. Here’s the last question! Q4) What are the solutions for the #CISO looking to use #AI and #humans in #cybersecurity? #AIChat pic.twitter.com/HxJ9iJMycX
— Nick Tang (@nickhtang) October 24, 2018
A4: Threat Intelligence is your friend. Using #AI to pull from worldwide feeds with a simple search will be able to seriously beef up your security strategy. Any sort of #NGOA is killer too. The hackers are acting fast, so we need to be responding fast. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
A4 (2): Being able to orchestrate and automate workflows and policies will assist with that. Ways that can augment your human team (especially with the talent shortage) is where you want to be looking to implement #AI into your security strategy. #AIChat
— Tricia A. Howard🦄🎭 (@TriciaKicksSaaS) October 24, 2018
I think the responsibility of a #CISO in regards to #AI and #ML is the same as current Cybersecurity practices. Enforce good practices that shrink the attack surface and set you up for success. AI is just another tool to help modernize and achieve that goal. #AIChat
— US Cybersecurity Mag (@USCyberMag) October 24, 2018
A4: Most of what we’ve discussed on #aichat today has centered on a unified approach between humans and #AI versus a silo approach. One cannot exist without the other.
— Adlib Software (@adlibsoftware) October 24, 2018
A4: AI/Automation should be used when knowing the problem you are trying to solve, which is a process. No point using AI on an inefficient process. That’s why you still do the hard work of modelling first (OODA works) before you implement. https://t.co/3P5igeDI6p #aichat
— Mark C Stafford (@StaffordCyber) October 24, 2018
How #AI Can Be Applied To #Cyberattackshttps://t.co/ft2V0LdD1O by @d0znpp via @Forbes#CyberSecurity
Cc @Shirastweet @digitalcloudgal @TriciaKicksSaaS @TamaraMcCleary pic.twitter.com/Hil3UACTg3— ipfconline (@ipfconline1) October 26, 2018