GM Joins Tesla and AI Investing

From Mahart that we're Innovation, Money and Power Collie in Silicon Valley, NBN.

This is Bloomberg Technology with Caroline Hyde and Ed Ludlow.

I'm Caroline hide a Bloomberg's World headquarters in New York, and I'm Ed Ludlow out here in San Francisco.

This is Bloomberg Technology coming up.

Ed will break down GM's decision to join Tesla's charging network.

This is the ev makerheads for a record setting rally.

Plus, we discuss all things AI, from its use cases in the field of law to investing in the space with venture capitalist Wesley.

Chat And for today's Bloomberg Big Take, we'll discuss how artificial intelligence is taking racial and gender stereotypes to new extremes.

But first let's check in on these markets.

Ed, we're kind of muted ahead of the Federal Reserve next week. This is a macro perspective on how much the Federal Reserve will pause, will they then hike in July. But at the moment, we've have seen the downward trajectory for the dollar. On the Weekles's in the s and P five hundred, managing to get into that ballmarket territory twenty percent off of it's October lows. We're currently clinging to gains only up two and a half. Now's that one hundred up higher, up to tens percent, some big

Pause this hiking cycle.

Now's some a little look at what's happening in terms of risk.

As said of.

Choice, we're currently seeing bitcoin having a pretty tortuous few days five day run and we're off by two point nine percent.

We're still clinging to that twenty six thousand there and thereabouts, But all.

Of this comes on the back of some significant regulatory headwinds. Look at that downward trajectory when we got the news on binance and of course some coinbas when it comes to what the SEC is doing. So notable that at the moment we're not seeing that move into textocs being reciprocated by moving in to bitcoin ed.

One big influence on this market is Tesla up five percent. It's trading in its highest level since early October. End of September, the news General Motors is adopting Tesla's North America charging standard. It is also having upside impact for GM, up two percent.

But look at the downside.

Movers Evego charge Point principally CCS standard offerings for charging networks really taking a hit. The Tesla name has momentum. Just quickly look at this terminal chart. We're up for eleventh straight session. That matches the streak of games we saw at the beginning of twenty twenty one, eleven days that ended in early twenty twenty one. We don't often get technical on this show, and I'm not showing it in this chart, but I think about relative strength index.

As we looked at this and realized we could really address one of the biggest issues that customers are telling us is, Hey, you know, I like evs, but if it's going to be my only vehicle, I need to know that there's a robust charging infrastructure. So I think this gives us a huge opportunity to do something that's better for customers and to drive it to be the standard.

Is this just the next domino to fall? Is this something much bigger? Let's get the details with Bloomberg's Detroit Bura G David Welch. David tell us how GM's going to pull this off.

Well, they're basically going to equip their cars starting in twenty twenty five with an import that can take Teslass chargers. Before then they'll have the customers, the owners of those vehicles. I have to use some kind of adapter so that they can use Testla's vehicles. But basically it gives these gives GM electric vehicle owners and forward for that matter, because they announce this deal is a similar deal previously gives them access to twelve thousand Tesla fast chargers and

Honest, David saying, another win for Tesla. How does this actually help the revenue of Tesla? How does this well help the business model of Tesla.

So, Tesla, they still are going to make all of their money or most of their money selling cars, but they do have this business of charging where they do make some revenue. It's not huge money. Estimates are that Tesla can make it anywhere from one and a half billion to three billion by twenty thirty and even more after that with all of these customers of other companies

David on that point about that Tesla's networks reach and maturity, We actually have this chart from Bloomberg and you Energy Finance which shows the Tesla fast charger network relative to ccs, and the chart tells the story, right, It is just so much bigger here in the United States. But the reason I bring up as well is the title of the Bloomberg NIF research. GM cruels back to Twitter and Tesla to save it's ev ambitions. You know GM better

Oh?

Look, first off, when Ford made the announcement, I went to GM and said are you following next? And you know they sort of hemmed in had I think they realized once Ford did that, they had to go in. And you know, Tesla and GM I've sort of traded barbs over the years, subtly at times, not so subtly at others. So I do think they didn't love having

You know, Caroline, this is clearly giving Tesla Antiema boost. But again, let's go back to those movies to the downside and some of the rivals out there in the network.

Yeah, Devid to that point, you mentioned EVgo, but also Pilot, there was a joint venture already announced by GM as you referenced. Who does this hurt perhaps more in the startup space and the areas that you know, I think of here in New York, the announcements that are being made of all the EV charging that's going to be put in the capacity, and it was helping players other than Tesla, who are the other battery operators, I mean the charging operators out there who perhaps stand to forfeit here.

Yeah, look, this is going to be tough on them, at least in the near term, I think medium long term they can adapt if Tussel's chargers are more reliable and I on a Cadillac Lyric are a four one fifty lightning electric vehicle, and I can use it six months from now because I get an adapter. Why wouldn't I do that if I know it's a better network. So they're going to lose potential revenue on this thing,

David Welch, as you say, GM aficionado, EV aficionado charger Officionado, We thank you so much for Detroit.

A New York lawyer told a judge he never meant to fool anybody when he filed a court brief full of phony legal precedents invented by chat GPT. The lawyer who faces punishment to the brief are the US district judge for leniency, claiming he had no idea the free artificial intelligence school could create fake case citations and court opinions. Generative AI tools promise to disrupt industries, but the technology also has many risks associated with it, so staying in

And Sami, I'm going to start with you.

Sure a month ago you came on this program and told Caroline and I, I have a billion dollars to invest in artificial intelligence yep, And I said, well, how are you going to do it? And you didn't fully answer the question. This goes some way to answering the question.

In the there has been an outpouring of interest, of course, by entrepreneurs around the world inbound coming in saying, hey, this is exactly these are the folks to come and help us launch our ideas, and this is what we knew would happen. Whenever you have a paradigm shift as profound as large language models and generative AI, the creative juices of all the entrepreneurs around the world get going and they try to solve novel problems in a novel way.

And that's what we've seen and even.

Up as one of the very first examples we've seen coming out of this allocation of that billion.

Dollars, Robbie, that creative juices that got flowing. We're actually started in twenty twenty, and I'm interested how much has generative AI changed the game for you?

How much?

And already you been iterating on making legal work in particular that much easier, and is sort of the momentum around chatchipt just helping with a cell cycle.

For sure, we've been experimenting with generative models since the founding of the company, and we also use discriminative models as well.

And the way you can think about.

What we do is we turn raw case files, typically medical reports and police reports, into AI generated legal documents for injury attorneys. These legal documents value what these personal injury cases are worth, enabling injury attorneys not only to save time in legal drafting, but also maximize the value of their claims and helping the many millions of Americans they get injured every year in the US.

Romie, what is your response to the story that we outlined at the beginning of this segment, a lawyer who now has to defend himself because he relied on fake legal precedence that chat GPT gave him. What is the risk for your company in that context?

I think it really comes down to how you're using generative models. The way our approach to generative AI is very different than a lot of other companies were strictly focused on one type of legal letter called the demand letter, and this is where the way we use generative models is in a.

Pretty confined way.

We feuded a handful of inputs and we're asking it to summarize it in pros.

Since we're not.

Scaring the web or asking for or a lot of different types of use cases for general models, the risk of hallucinations is significantly less. And separately, our demands are acuated by a team of legal professionals. We've done thousands of these cases, and our clients haven't seen those types of results of hallucinations impacting their business.

Submitted to that end, you know how, was even not attractive because it is such a specific application of generative AI and AI in general. Are you looking for those specific areas, particularly in healthcare, particularly in legal work, or are you seeing more broad applications that excite you to.

Certainly all the above, although I would say even up is a perfect company for a Bessemer venture partners to invest in. Bestmer has for the last ten or fifteen years been I would argue, the leading venture capital firm investing in vertical SaaS companies. So whether it's Toast in the restaurant industry, Service Titan for plumbers and electricians, pro Core and construction Shopify for e commerce, been focused on investing in companies that solve the particular problem set of

What was doing this deal? Like what does the cap table look like at the moment as well? Like, yes, you're on this, so it's Spain. So are many others. Yeah, you know who's convincing? Who you convincing? Rammy, Remy convincing you. It is a very competitive environment. Ronny is an incredible entrepreneur.

There were he could tell you how many venture capital firms he had interest in in it. We were privileged that he chose to work with us.

Very polite answer, and and we have many friends.

The folks of being capital ventures are friends of ours from.

A decade long. So you're seeing more and more of that.

I think partnering go on on, but in spirit of serving, and I think you'll always hear that from from Investmer. We're in service of our entrepreneurs and in service of those companies.

So Romie, I'll give you the opportunity to give your perspective on that.

I would note, you know, do not pay.

Josh Brown has been on the show with Carolina and I they do something quite similar as well. So how did you stand out to all these bench catalysts that were fighting over you?

For sure?

And our approach was just to do one type of document and more affordably, more accurately and faster than a human can on their own. And knock on wood, we've had some pretty good traction. It speaks volumes, just at the value proposition, being able to deliver it to injury attorneys and again, ultimately their clients were injured us. What is most important is finding a partner that is online in that mission, and I think Samor sold them a

Robbie, you just mentioned there quicker than a human can. Is this going to reduce jobs?

We have not seen that.

Ultimately, if you think about personal injury, you're talking about some of the most vonderable times in individuals' lives. You're living a normal life, then all of a sudden, something really had happened to you. I can't imagine a human not being there to help help these folks share through the personal injury claims process. From what we've seen from our clients, they're spending last time on document review and document drafting, and they're spending more times with our clients.

Ami Caravaba, thank you very much. Indeed, even up CEO Sama, great to have some time with you as well. Sammy, are lacking a partner at best in the venture, but partners, that's so much for joining us. That was a really interesting roundtable about all things generative applications. Meanwhile, coming up, we're on the fallout actually from the Securities and Exchange Commissions lawsuit against Finance. What the crypto exchange platform is telling its customers.

From New York. From San Francisco, Disi Bloomberg.

It is time for talking tech.

First up, Taiwan semiconductor Manufacturing, seeing a dip in revenue for a third straight month. Sales at the world's largest supplier have made to order. Chip slid five percent, which indicates that the sector's slump has yet to bottom. TSMC executive has signaled a gradual recovery in the second half of the year and a resumption of growth in twenty twenty four. And Binance US is telling its customers to take action in order to protect their assets following the

End quote plus.

Crypto firms are reshaping the landscape with its come to a banking system in the US. Firms are turning to smaller regional lenders to open accounts. Swiss and Asian banks were also playing a bigger role as to where assets are placed. The changes come amid increased scrutiny from regulators against that industry.

Caroline, in fact, cryptoregulation and all that was being discussed in a venture capital conversation I had yesterday, Ed, I got a chance to sit down with Rebeca Cayden of Union Square Ventures based right here in New York. Dina Shaker of Lux Capital as well, had flown in from the West Coast where you are for the Bloomberg invest conference, and actually when it turned to AI, they expressed some optimism there the role in the shift that we're seeing in fintech and healthcare.

Just ticul listen, I'm an optimist. I think I'm partly an optimist by trade and overall, but particularly here. I think the opportunities set to unmark and the efficiencies to create, and the skills and abilities that we can't think of yet are going to outweigh the risks which are there and are are relevant, but are going to be manageable.

Okay, I like that they're manageable. Dan, Do you have that same optimism that ultimately all the ringing of hands, worries about job implications bias will ultimately be outweighed by the positives and productivity and how it's going to help your particular area of expertise, like.

Healthcare, absolutely.

I mean, we've been investing in AI for years, so if certainly there is a lot that's new now, it does feel like a paradigm shift. But we were early investors in Hugging Face and Runway. In the healthcare space, we actually sit on a board together of a company called a Life Health, which is using AI to transform fertility care.

There is still so much.

More and we're incredibly excited about the potential, of course, notwithstanding the regulatory environment that we're in right now and making sure that everything is done safely and also without bias.

Go to regulation in a moment.

But to the point of where AI is really disrupting?

Are you where's your rain space app?

Is it more inbound people wanting money, people cooling themselves, ultificial intelligence companies you try to.

Sort wheat from chaff?

Or is it how do I ensure the port photio Companies I have that aren't truly built on AI are ready for that disruption?

I mean, I think it's a little of both.

Obviously, as a consumer I love using it for bedtime stories for my kids and you know, recipes and so on. But as an investor, we're looking at what the really big platform shifts are going to be and I think any tech company right now needs to take a really close and deep look at how they're using AI to make themselves more efficient, particularly in well, frankly, in every field. But I would say when it comes to healthcare, there's so much opportunity, so much latency, so much bureaucracy.

But it's a very different world when you're still.

Talking about fax machines in the same sentence.

As chat gpt Well said, And what about the fintech space, Rebecca, and in this is an area that you'll particularly got some expertise in as well.

Across the board as well as healthcare.

Some have said maybe fintech isn't the first era of adoption for AI.

What would you make of that?

I think that the enco across industries and healthcare and fintech and education have a big opportunity here, less because of their industry specific nature that there's some of that, and more than the idea that where there's rote tasks or efficiencies to be created by people, it can be a very powerful tool to create that. And that's cost savings, that's time savings, that's adding value into the stacks that's already there, and you know, you asked about kind of

It was such a great conversation. You can go online to see more of it.

Rebecca Cayden of Union Square Ventures, Tina Shakir of Lux Capital.

Welcome back to Bloomberg Technology. I'm Med Ludlow in San Francisco.

And I'm Caroline Hyde in New York.

Let's check in on these markets because nasdak treading some water today. We're up to eleven points, but remember it had hit the higher since April twenty twenty two at one point, so the SMP get into a ballmarket technical ball market, twenty percent off.

Of its lows.

We're still worried, tentative about what happens with the Federal Reserve next week here in the United States, So at the moment, maybe just a little bit of caution as we end up the trading week basically flat on the week for the NASDA two year yield, though up seven basis points as we think about where in which the Federal Reserve has to go, how it has to tackle interest rates, whether or not that's really priced in into the front end of the curve.

So two year yield's just pushing up higher.

Bitcoin currently down by sixteen percent torrid week, int of course the regulatory overhang for crypto in general, but actually managed to be relatively resolute in the face of that. Twenty six thousands is where we're currently trade. Moving on, and let's go into some individual movers, because there are some.

Big moves going on.

I'm looking at its Heather, of course, up almost four percent on the day as they sign that deal with gm as look really to be leveraging the power of their charging that they have at the United States at the moment. But Adobe is on the higher side, up more than four percent, going overweight in some analysts or really thinking that this is a company to be leveraging, whereas we look at generative AIAI applications.

Interesting that Peloton's at more than four percent. Who knows why.

Big volume at one point up nineteen percent on the day at the moment, Bluemode not reporting any real reason and catalysts into the Surgeon volume and actually the Surgeon shares. But we're coming down to about four percent, but one standard deviation move compared to their twenty day moving average. Blue Apron as well. Look, it's a penny stock. Basically, we're seeing it only what fifty million dollars worth in market cap.

It used to be worth one hundred million, in.

Excess of perhaps sixty percent on the day. This is a pay off some debts away from one to be watching as well. But we want to be turning back from public markets to private markets at the moment. Ed and a business planning platform Pigment just reported in an eighty eight million dollars Series C funding round led by Iconic Growth as well as Felix Capital coming in on the deal as well. Let's bring in the co CEO co founder Eleanor Crespo now who's joining.

Us from France. But the focus that is funding around is look.

Eleanor, you're going to be beefing up the business, going to be pushing more globally here in the United States as well.

What services do you offer?

Thank you very much for having me so.

First of all, just to reset this scene, to explain for everybody who we are. Pigment is a business planning platform that Eulu's business leader no materally signe finance if you are a CEO, the CRO or the HRVP of your company, to take better, faster and smarter decisions. The other point of Pigment is to quickly adapt to change.

Current macroeconomic environment.

So to give you a concrete example of what we do of today, as you can see or you see, the macroeconomic environment is what it is. You have inflation raising a different paces in different countries, and so any business leader needs to be able to answer in the very first fashion, for instance, if inflation raises even more in North America, how that will impact my hiring, how that will impact my growth margin.

How that will affect my self process?

And ultimately Stigment is here to offer you to help you answer better these questions.

What's so fascinating is the problemly you're solving, but also your own background, the fact that you were a vcy yourself before an index, the fact that you were working at Google for a long time. In terms of strategy, what was the funding environment like for you? How easy or hard was it to sell real vision and the need for capital take a macroeconomic head wind moment, they were just painting for us.

Yeah, So actually we were not raising these rounds and what happened was very different. So we still had the money from the past two runs, pretty much everything in the bank. So here actually we got a lot of interest and that came from a couple of reasons. First one being that we are must have platform in the current economic environment.

Everybody, every customer.

Across all industries need to plan on the two like hours. That was the first thing, and the second thing was I think the type of customers we managed to sign across all industries. So we managed to actually get on board and serve the most incredible customers in the likes of Figma, of Mirror, of air Table, Mozilla, Kelvin, Klein arp et cetera. And really what happened with them is that not only they showed to investors that were they were using us on.

A daily basis, but that also they loved us.

And what happened really with this round is that the stores actually got some word of mass from all of these customers that you know, this product was a total revolution in the way they plan. And we actually got this from preempted biobsisis leading firms that you mentioned iconically dig the round alongside Felix, Capital IVPN, Meritech participating. And really the reason was I think the attraction that we got not only that, but we also managed to go

Year and that elent know. Eleanor.

I want to jump in and just ask you because I want to understand some of how you manage that. First of all, you have runway right, so you raised eighty eight million dollars in this series C two hundred and fifty million today, where did this round value your company?

So, actually, we are not discussing valuation for a very good reason is that, first, it's.

Not a focus at all. What we are very proud of at.

Pigment is our customer base or partners and the fact that we had the tremend destruction in the past year. You know, we grew our revenue by seven x last year and in north summer tenx in you know, the number of users that are using Pigments. Really, I think the valuation is not a metrix that we want to follow because we think it's not really the real value.

So it's in response to growth, right, tenx, growth ten x in the user base last year. How do you deploy the capital to respond to that? You having to manage headcount, infrastructure costs.

Yes, So actually I think we are going to do two major investments this year, so obvious see one of them, which is a continued investment, and the most important one is in the product. So we are the most innovative platform out there by far, and we want to remain the most innovative platform. So there are a lot of very exciting things coming in the roadmap, AI being one of them. Of course, I'm very happy to tell you

What is important is to be able to bring the most tremendous customer experiense to all of our customers. So these are the main two areas where we are going to completitles this year.

Caroline, I cannot resist the temptation. We have a startup in the software space raising money. Let's go there and talk about AI.

I mean, it got to you. Just let us there.

I don't know how, I mean artificial intelligence and just looking, we saw some video of how you interact with your product and it only looks like it sort of generator AI and predictive AI in some way.

How are you folding that more into the business model?

Yes, so, actually AI is already something that obviously we had thought about. It seemed the beginning first of Fuller with my co founder Romanically, we started to integrate an AI roonmap from day one and as of today we see we have several new features in better with the partnership with Chagpt.

And for us, that's what is critical about the way we integrate AI within the platform is to make.

The most intreitive customer experience for pretty much any business user on the platform. Again, our goal is to have literally any business user within a specific company to use Pigment.

So for us, it's all about the simplicity of the platform.

I can give you a concrete example around natural language that you can use with in Pigment to ask question such as what was my revenue in North America asture and that you can ask in natural language and the platform will answer. So for us, the idea is really to democratize even more the access to the platform.

Pigment co CEO co founder Eleanor Crespo, French entrepreneur, French startup coming.

To the US market Carroy, Thank you for joining the show.

Now coming up All Things AI investing more with FPV Ventures Wesley chan As.

Next, this is Plameberg.

The hype is absolutely remarkable and I've never seen anything quite like it. AI has been having an impact for decades and this stuff is into brand new.

If I'm right on AI and the impact on it, I mean, it's already making the top coder seven eight times seventy eight times more productive than they were five months ago. If it's as big as I think it is in Nvidia, is something we're going to want to own for at least two or three years, not for ten months.

We want to engage with the regulators to say, what's the next generation of AI you going to do that's going to make it even more effective. And therefore maybe it's not totally explainable, but you understand that we're not using it, we're using it for the right purpose, and so changing the regulatory I would say process around the use of AI is going to be an important except for every industry.

More software, That's how I would see it. And we've been bringing more software into finance for a really long time, and there's been all kinds of breakthroughs, and there's been all kinds of problems. It is statistical pattern matching. It's extremely powerful and it's interesting, but I do not see AI achieving what some would call the holy grail. Everybody wants to know what's the S and P going to be in six months?

Can you tell me?

And I can and neither can the AIS.

Ah some hype, maybe some reality there.

So speakers from Albertneberg invest some it this week giving their thoughts on all things AI. The space of course that we can't get enough of. Let's stick with it for our VC Spotlight. Very pleased to welcome Wesley Chan for more. Here's of course the co founder, management partner and FPV Ventures. And I'm sure you have a view or two whether we're hyped in terms of AI exuberance valuations.

Even know, it's.

Fascinating, right, I've watched this cycle happen so much many times.

You know, I've been an.

Investor for fifteen years and this happened when mobile phones first came out. There was a hype there and you know there were all these funds started on mobile and today you know everything is mobile, right, Like can you imagine trying to book an uber on your web browser. This doesn't happen you do it on your phone. And that hype sort of happened in two thousand and went away because we had everybody incorporating uber I mean sorry

Okay, So to take the long term view, managed to sort of move out of the exuberants.

And the headlines.

How are you picking which companies in this current cohort you want to be backing, because I know you're all about the founder, but how are you ensuring that the photio companies you have are resilient to the disruption and how the founders that you back right here, right now are the right ones?

Again, we take our lessons from history.

In twenty ten When I started at Google Ventures, we invest in a lot of mobile companies, right. You know, I was an investor in a company called Parse, which wound up selling the Facebook, and you know that Facebook wound up shutting it down. It made some money, the founder did well, but we those mobile companies aren't around anymore because again it starts being incorporated as table stakes into every product. I think we're going to see some of that with AI again. Right, this is a very

Rushing in AI.

Some firms are creating AI funds just specifically for looking at AI, and we are not investing in AI companies for the sake of AI. We're investing in companies because there's going to be great business models that would be created by the AI, and we don't even know what those are yet. There were all these mobile infrastructure companies in two thousand and ten when that happened. All those no longer exists. There's very few of them. The companies that wound up doing super well in that market had

I think we're gonna see the same on AI.

There's gonna be new companies that will will change the landscape of how we do business, how we do work, how we buy products, how we book cars, how we book flights, and we're looking for those companies where that business model will be changed by AI rather than AI. For the sake of it, right, there's a lot of hype going into it, a lot of money going into it.

Wesley, it's been almost a year to the day since you were on Bloomberg Technology because you'd raised and launched your four hundred and fifty million dollar funds.

YEP.

Based on what you just said, have you not pivoted at AOL in the first six months of this year to respond to what's happening around you in terms of strategy and where you deploy capital.

We are investing in mission driven founders that have a unique insight on how to change the world. When Larry and I spent my first fifteen years of my career at Google, I joined very very early. There were very few people at Google. When I joined, it was Sergei Brin's chief of staff, and Sergey is one of the co founders of Google. Google is an AI company today. Back then, nobody saw Google as.

An AI company.

They saw it as an ADS company the search engine. But Google has been using AI for fifteen years. We're just investing in founders that get what new business models get enabled because of AI. And We've been doing this for the last year, and I've been doing this in my career when I was at Google Ventures, in my previous fund for the last fifteen years. Some of our best companies are enabled by AI. Right, there's business models certainty,

They're using AI to figure that out.

There's business model certainty in these businesses.

Right.

If it's a drug company and the AI helps them get to a drug faster, then we know that the drug is very valuable and the market values that. So those are the type of companies we're investing in. They're run by great founders, and the founders truly get that AI is helping them get to the value faster rather than just creating a new a AI.

Wesley.

Those are the types of companies, and I actually am really interested in the mechanics of how deal making has changed over the last twelve months. So you launched with four hundred and fifty million last year because of the inbound and activity around AI, are you having to write larger checks or write checks with more frequency to keep up with the energy of your industry.

Well, we've never, ever, ever invested into the hype when there were grocery delivery companies where people were putting billions of dollars. I remember betting pitched five with these grocery delivery companies, you know, in my place in New York. Every week I was getting a new postcard for thirty percent off some fifteen minute grocery delivery company, and at one point there were fifteen of them. Right, We didn't invest in any of them. Same with the scooters. We

In fact that you know, every time we.

Do a deal, there's probably like one or two other investors, you know, looking at that space because everybody's chasing these over hyped AI companies where we don't know which one's the winter. So we're very very contraran in how we look at companies. We do AI, but we don't do AI for the sake of AI. Just like in twenty ten, all the people that invested in mobile for the sake

Fpv A Ventures, co founder of managing partner WeSC Chann, We fit a lot in there.

Thank you very much for your time.

By Dance, the parent company of TikTok, is testing its own AI powered chatbot, joining China's AI Race. Not much is known about Grace, the project's code name, but Bloomberg sources say it's being tested internally among by Dance employees. It's worth noting that TikTok is also experimenting with its Taco chatbot, which appears as an instant messenger on the app. Perpetuating harmful stereotypes and biases is one of humanity's biggest ills, but a Bloomberg analysis of a text to image tool

Okay, so Bloomberg.

Uses Stable Diffusion, It gives commands and generates thousands of images.

Talk us through what you did and why.

Yes, So we basically carried out an experiment with Stable Diffusion, which is one of the biggest open source platforms for AI generated images, and being open source makes it easy to also analyze and experiment with, and we wanted to understand how deeply ingrained.

Biases might be in this technology.

So we asked to create thousands of images of workers for fourteen jobs and also different criminalized categories, and then we analyzed the results and you know what we found was a really systemic pattern of racial and gender that doesn't just replicate stereotypes, but it actually makes them worse. It stretches them to extremes worse than those found in

It's truly depressing, and I suppose ultimately the problem here is society and stability AI and Stable Diffusion is just emphasizing that even more when you're coming to your reporting, when you're looking at the outcomes. How much was their discussion about the fixes here? Is it about auditing some of the algorithms, is there a way of ensuring that we don't or is it about the prompting that's being put in.

Well, one of the important things is about transparency. So it is a good thing that Stable Diffusion is open source because that enables people and you know, researchers to actually improve the model and make it better. So that's in a way it can be better than closed source models like you know DAI for example, where we don't really have that much knowledge into how they're being improved and by how much.

So yeah, yeah, sorry, let me just jump in real quick, and I would just point out that a spokesperson spinit Ai points out that all AI models have this inherent bias of them, but their point is that by open sourcing it, Caroline, that they can work towards overcoming that rather than close or non open source models.

I wish we could discuss this day and day out, and Leo, I'm hoping you can come back and join us a little bit more on but more imminently.

Go read his work.

It's with Dinabas as well. Leo Nicoletti, thank you for staying late for us in Italy. It's absolutely fascinating peace and ultimately a bit depressing as it stands, but we don't want to end.

On a low note. Ed because the time is nigh.

We're wrapping up. It's the edition of Bloomberg Technology wrapped up, but it can.

Get more incredible weak and I've forgotten that it starts with Apple and WWDC. You can recap on the podcast wherever you get yours, Apple, Spotify, iHeart and Don Bloomberg from New York from SF.

This is Bloomberg