As AI of All Things (miscalled) dominates the world’s biggest security events, we round up some of the most touted use cases and applications.
Well, we have this ChatGPT layered on top of the AI. Not really, but even the practitioners responsible for some of the most impressive machine learning (ML)-based products don’t always stick to the basic terminology of their field. . Expertise…
In RSAC, marketing and economic considerations, let alone the granularity of basic academic distinctions, tend to be superseded, and all the rest of the supporting ecosystem protects and implements AI/ML. , built to manage. This is not easy. .
To be able to answer questions like “What is love?”, systems like GPT collect different data points from many sources and combine them so that they can be used loosely. increase. Here are some of the applications her AI/ML rep at RSAC has asked for help with.
- Are job seekers legitimate and telling the truth? Sorting through social media chaos and reconstructing dossiers that weigh and contrast candidates’ glowing self-assessments is an acceptable time-sensitive HR department. Not an option as we struggle to scrutinize the mass of résumés that arrive at Shinbun. You can pass something like a short list to the manager. Of course, one has to wonder about the dangers of bias in machine learning models as they learn given biased input data, but this, while imperfect, is more likely than human-initiated text searches. can also be a great and useful tool. .
- Has your company’s development environment been infiltrated by a malicious person through a third party? A practical way to monitor all development tool chains in real time and expose them to potential hacked code issues there is no.
- Are deepfakes detectable, and how do you know if you’re looking at a deepfake? I was. The Real He asked the audience if they could tell the difference, and the answer was “almost, if ever.” So if the “CEO” asks someone to send a wire transfer, can you trust it if you watch the video and hear the audio? ML hopes to help you find out. But since the CEO tends to be in public, it’s easier and much better to train a deep fake of him from real audio clips and video clips.
- What happens to privacy in an AI world? Italy recently cracked down on ChatGPT usage due to privacy concerns. One RSAC startup used an interesting coding technique to provide a way to keep the data interacting with ML models private. This is just one attempt at a much larger set of challenges inherent in the large language models that form the basis of well-trained ML models.
- Are you building insecure code in an ever-changing threat landscape? There are many new coding practices that are proving insecure even if your tool chain isn’t compromised. Especially when it comes to integrating cloud properties with mashups. Correcting code based on such insights driven by ML can be critical to avoid deploying code with built-in security.
In an environment where GPT consoles are almost unnoticed and rudely sprayed to the masses, and people see the power of early models, it’s easy to imagine the fear and uncertainty about how creepy they might be. There is certainly a backlash to try and curb technology before it does too much damage, but what exactly does that mean?
Powerful tools require strong guards against cheating, but that doesn’t mean they’re useless. Somewhere in technology there is a moral imperative built into it that needs to be sorted out in this context. Meanwhile, I turn to one of hers on the console and ask, “What is love?”
