Could AI manage your portfolio better than a human?

As AI seeps into all corners of the market, Qraft Technologies is using AI to select stocks and rebalance its ETFs. Qraft Technologies head of AI ETFs Weldon Rice joins Catalysts to discuss how Qraft utilizes AI and how AI-enhanced ETFs compare to traditional ETFs.

Rice emphasizes that artificial intelligence is purely used to construct portfolios. However, he notes that humans construct the model itself: "We have experts here in-house that are helping to drive what data points are made and what types of models are being used and those sorts of things as we research more and more." He explains that once the AI model is created, it uses new monthly data to make investment decisions.

He points to a recent breakthrough in Qraft's partnership with LG AI Research, where the AI model can produce reports telling investors why it decided to pick certain stocks. "We think this is really the next step in AI innovation for investment decisions with AI," he explains. He adds, "The way that artificial intelligence looks at the market is very different from a human one. We're just looking at it purely from the data perspective." Rice explains that these models provide a non-linear perspective on the market that humans are not usually able to see, giving Qraft's ETFs a unique advantage.

For more expert insight and the latest market action, click here to watch this full episode of Catalysts.

This post was written by Melanie Riehl

Video Transcript

So I'm, I'm curious because you said that it's an A I driven ETF so you're strictly relying on A I technology to make these types of decisions, I guess.

How much is there this human element of uh of the human involvement?

And, and is there any rationale, I guess behind some of the moves that have been made here over the last several weeks with your funds?

Absolutely.

So, uh in terms of how we construct our portfolios, you're absolutely right.

We use artificial intelligence uh purely.

So, uh now, obviously humans are at the, you know, construction of the model.

So we're, you know, we have experts here in, in house that are helping to drive, you know, what data points are, are, are made and what types of models are being used and, and those sorts of things as we, as we research more and more.

But once the model is made and trained, it just takes in new data every month and then based on its training, it makes its decisions and we don't uh you know, change those.

So, uh yeah, when it, when it came through the NVIDIA, you know, I dropped down that much, you know, there was no human that said, oh no, that's the wrong way to go.

Uh We need to, we need to change that uh in terms of the rationale uh behind it, you know, uh that actually is one of the, the bigger challenges of, of artificial intelligence is that it is somewhat of a black box because we are, you know, using these deep learning models that are looking at very complex relationships, you know, uh relationships of relationships of data.

And so it becomes a little bit difficult to uh you know, understand that.

But uh one kind of breakthrough we've seen recently, uh not exactly in QR FT yet.

Uh but the recent fund that we launched in November with LG I research, um that fund has an element of using large language models, which is basically like CE PT.

And so now we've been able to uh LG I research uh has actually been able to produce a report uh looking at why the, the A I model picked the stocks that it did.

So, uh it's actually quite interesting.

Uh We're actually just now starting to put that up on our, on our website for investors to go look at.

And we think this is really a, you know, really the next step in, in A I innovation for, you know, uh investment decisions with A I.

What's the math that the A I is using, that investors are not using to suss out NVIDIA.

So I, I think that, you know, the way that artificial intelligence looks at the market is very different from a human.

Uh one, we're just looking at it purely from the data perspective.

But the way they were able to look at, you know, generally in the past, when we look at either fundamental managers or when we look at quant traditional quant managers, uh the the models ty typically are quite linear.

And so what we're able to do with artificial intelligence is as again, we're looking at the relationship, the data, but we're looking at the relation the relationships to the data.

And we get this kind of nonlinear look at the market uh that a human is not able to see.

And I think that's what makes this so unique.

And I think this is what is, is really promising for investors is that you're getting a very different look at the market that in your portfolio than you would from a, you know, traditional human active manager uh out there.

All right, Waldon Rice craft technologies, head of A I ETF very interesting.

Thanks so much for joining us.