Can AI Read Your Mind

Tom's Guide

MindPortal aims to revolutionize communication by using AI models to convert brain signals into text, potentially paving the way for telepathy and seamless interaction using ChatGPT. Imagine thought-to-text communication! However, as our AI Editor Ryan Morrison discovered, this AI mind reader technology still has a long journey ahead. Find out why the experiment didn’t quite hit the mark and what challenges lie ahead for this ambitious startup.

Transcript

00:00Have you ever thought tech is listening to your thoughts? Well, that's exactly what I'm here at

00:04Startup Mind Portal to have done. They've created a new technology that can read your mind and send

00:10it to ChatGPT so you can have a conversation without having to type or speak. Wish me luck.

00:19Hi. Hi Ryan, good to see you. Hi, good to see you. So tell me a little bit about Mind Portal.

00:24What is it? How did the idea come about? Mind Portal is a human AI interaction company

00:30and it was founded after a year of me taking psychedelic substances. During those abstract

00:38thinking sessions, a lot of visions came out. So during those sessions, what I set as kind of

00:43the thinking goal of those sessions is the future of humanity, how to make an impact, etc. But really

00:50the premise of Mind Portal is we want to explore the nature of human AI interaction.

00:54Okay. How can you make a future that symbiotic, which biologically is, win-win?

00:59We're going to look at a demo today where we can chat to ChatGPT using our brain.

01:03Yes. So this is the world's first demonstration, like never before has this been done. Can you bridge

01:09the gap between the human and their thoughts, which is the most intimate form of communication,

01:14what you think, what you feel, and an AI agent that can understand and respond to those thoughts.

01:19All right, should we try it? Yes, let's do it. Tell me what we're looking at now. So you've got

01:25something called an EFNIR system, which is able to record brain data, optical brain data, based on

01:31the blood flow happening in Ed's brain. When Ed imagines language that obviously activates different

01:37parts of the brain, it's that activation that's being picked up in real time. It's also the first

01:43demonstration of communicating with ChatGPT as well. Using just your mind. Exactly. You'll think a

01:50thought, the sentence is classified, but that classified sentence then becomes an input into

01:55ChatGPT to then respond to you. Okay. We're going to have a look at it. What's the goal? What do you

02:00want to do with this? Where do you see it going? This system, if you were to spend time building it

02:06towards commercialization, you could scale it in a multitude different ways. Okay. So number one is you

02:11could scale the number of sentences and then of course you scale the accuracy. What are we doing

02:16first? So if you'd like to pick a sentence. All right, well, let's go for Venus. I'm a space buff.

02:22Yep. I'm going to imagine this. Well, I'll read it out loud then, because if you read it out loud,

02:26then it raises the question of, is it just taking it from your voice? So you're going to think in your

02:31mind, if I were on Venus, I would be in a world of extremes. The pressure would feel like being a kilometer

02:37underwater, crushing you from all sides. The air is a corrosive nightmare capable of

02:41dissolving metal and forget about rain. It's sulfuric. So you're thinking that.

02:46Yep. That's a long sentence to imagine. It is. So you can't just imagine the visual

02:51of being on Venus. You've got to imagine the actual words in that sentence. Currently,

02:55yeah, we're trying to extract the semantics from that. All right, well, let's go. Let's see how that

03:00happens. You're going to think those words and hopefully the chat GPT will respond.

03:06So you've sent that off as the prompt to your decoder.

03:13So now the decoder basically is taking the brain data and trying to identify which of the sentences

03:18he was thinking. Okay. And then that's over there is outputting the sentence. So you got it wrong.

03:26So this is the restaurant sentence. Okay. So it was sentence number two, I think.

03:30And this is the brain data that went, that as he was thinking that sentence went into the system.

03:35Yeah, we can do another one and we can show you basically how this progresses as it's imagining.

03:40What's the one that works more often than not? If we're sticking to probability.

03:46Let's try it. Let's try again. Let's see. Let's see if it works.

03:54And right. Okay. So this time he's had a chat with mum on the phone,

03:58but it is showing that you can have this conversation. It's just a case of scale.

04:03Correct. So with enough data, with a larger model, we hypothesize as we've seen in AI or with any

04:11breakthroughs in their infancy, the accuracy would improve. And then you'd start to increasingly have

04:17the conversation you want without the incorrect inputs. In essence, what's happening each time,

04:21there's an incorrect input and then there's correct ones. And correct ones are happening enough times

04:25for us to know this works. Okay. Scaling data and scaling the model is let's get it to work more and

04:31more times with reduced error in essence. Should we try it one more time?

04:37At least we've seen them all now. No, it should be stressed. This is very early stage. We're

04:42looking at a sort of research preview of a technology that with enough scale will improve

04:48potentially exponentially. Where do we go? What am I going to be able to go into a supermarket

04:55and look at a product and say in my head to my AI, can I eat this and have the AI pick up my thoughts

05:04and respond? Can we get to that point? Yeah, I think we can. And the reason being is we've seen

05:10this again and again in AI, as you increase the amount of data, as you increase the model size, you get

05:17better performance. So the time constraint, honestly, there is how many people and financial constraint

05:23is around how many people can you collect brain data from? Because unlike going online or using

05:29written sources, which are easy sources of data to acquire, this is a bit more of a tricky in the

05:34current paradigm. I've done some back of the napkin calculations, just for fun. And it's not as

05:39expensive as you might think. All right. And it doesn't take as long as you might think. I think for,

05:44you know, under $50 million, which is, you know, in the venture capital world or in the world,

05:49this is not a change. Yeah, exactly. You could have in six months time operating 100, 200 different

05:55headgear caps. Yeah. People coming in in batches and have thousands of people going through. Now,

06:01of course, my cursory kind of calculations assumed a threshold you'd need to reach because we don't

06:07know how much data will confer. So let's assume you reach that threshold, you get the funding.

06:13And in a year's time, you've done all the data, you've crunched all the data, your model's working.

06:17I can go out and buy a baseball cap and talk to my AI without having to speak out loud. Can I talk

06:22to someone else wearing the same baseball cap? And we were going to have full telepathy

06:26with the AI as a translator. So the answer to that is, if you scale the data and if you scale the model,

06:31and if you integrate it into a cap wearable, then yes, theoretically, it should work. It should work.

06:36There's no reason why that shouldn't work. So we can have telepathy.

06:39You could have telepathy. There's nothing. And that's what we were setting out to prove. So for

06:43example, I could wear headgear, think of a sentence such as, how are you today? Yeah.

06:50That could be then sent through an AI model that takes the text and translates it into a voice

06:55Yeah. And puts it into your ear as an AirPod. Okay. Through an AirPod. And you can hear me

07:00And you can just respond. They can respond with their thoughts. And you can respond back to my AirPod.

07:02So in theory, we're having a telepathic conversation. Neither of us are speaking.

07:06Yeah.

07:06But we're using pre-trained sentences to have a back and forth dialogue, which we're both hearing.

07:11And now you've got AI models that can take a bit of your own voice.

07:15And it can sound like you. And put it into the sound.

07:17So it would sound like Ryan when I'm hearing. It would sound like me when I'm talking to you.

07:20That raises an interesting point because that would potentially give the voiceless a voice.

07:24Because you could use a text-to-speech engine based on that.

07:30And their thoughts could go directly to the voice engine rather than having to type it out.

07:34Exactly.

07:34Well, that was a lot of fun. It didn't work as expected, but it's an early research preview.

07:39This isn't a product they're going to be putting on the market tomorrow.

07:42However, it did give us a really interesting insight on what we might be using

07:46and how we might be interacting with AI and each other in the next few years.

07:51And I really hope it works because I do not want to be standing in the supermarket

07:57talking to myself when I'm just having a conversation with my AI.

08:01Fingers crossed.

08:02If you want to find out more about what's going on in the world of AI,

08:05find me on Tom's Guide or follow our socials at Tom's Guide.

Category

Transcript

Be the first to comment

Recommended