AI Trust Depends on Data Control and Safeguards (Full Transcript)

[00:00:00] Speaker 1: AI is possibly the most powerful tool any of us have had access to. And that's quite exciting, but also a little bit scary. And I think there's another side of the AI conversation we're not talking about enough. And that's trust. Because we're no longer just trusting AI to write emails for us anymore. We're giving it documents, our work, our conversations, our personal information. And some of these systems can now remember context, analyze what we're saying, make decisions, and act on our behalf. So naturally, people are starting to ask, how safe is all of this? And honestly, I think that's a completely reasonable question. Because when you give an AI tool your 45-minute recording, you're no longer just uploading a file. You're potentially giving it someone's name, their opinions. You'll come with a strategy, pricing, and confidential information. Basically, an entire conversation. So I want to use Happyscribe as an example, because this is a product where it really matters. Let's start off with the obvious one. Is my data being used to train AI? With Happyscribe, model training data is optional. You can use the platform without opting into that use. And the setting can be controlled from your account. And I think that's a pretty important distinction. Because when you're dealing with AI, I don't think you should simply just trust it. You should be able to understand what happens to your data, and then be able to control it. The next question is where your information actually lives. Happyscribe stores data in the EU. And the platform uses encryption both when data is being transferred and when it's stored. It's also SFC2 Type 2 certified and built around GDPR requirements. And I know that sounded like a bunch of boring security terms, but that's kind of the point. When you're uploading private information to an AI, boring is kind of good. Because you want security processes. You want access controls. And you want encryptions. Because you want to know where your data is, not just a cool AI demo. And then there's another concern I think people have started to notice. The AI meeting bot. Because to some people, there's something quite strange about joining a call and then seeing an AI bot join the server. Whereas with Happyscribe, you can use the meeting notetaker when you want it. But also record without putting a bot into the meeting at all. You can just use the browser recorder or simply record from your phone. So if you're sitting across the table from someone or interviewing someone, you can capture it without needing an AI bot into the conversation. Now this is another point I think people overlook. It's also about who inside an organisation can see something. Happyscribe has permissions and private workspace controls. So you can control who has access to the recordings and transcripts. Because not every conversation should be visible to everyone on your team. Now this next point is probably my favourite part. Because I don't think we should pretend AI is perfect. Because if you need a transcript where accuracy really matters, Happyscribe also has a human transcription and proofreading service. So you have an interesting kind of safety net there. AI when you want speed and human proofreading when you need that extra layer of accuracy. And I think that's a much healthier way of thinking about AI. Because you don't have to blindly trust it. You use it where it makes sense and you control it where it matters. So the question is, is AI safe to use? I don't think the answer is as simple as yes or no. The better question is, what does an AI company do with your data? Where's it stored? Is it being used for training? Who can access it? Can you control it? And what happens when you need something more than just an AI generated answer? Because AI isn't going away, we're going to even need more of it. The important thing is making sure that when AI gets more powerful, we don't give up control of the information that makes it powerful in the first place.