Microsoft reaps its AI rewards. Its customers? Not so much

Big Tech earnings are starting to roll in, with Microsoft and Alphabet reporting earlier this week and Meta and Amazon having their turn today. I get into Google toward the end of today’s newsletter, but we have to start with Microsoft. The company recorded its highest profit growth in more than two years, largely on the back of AI-fueled growth in its Azure cloud division. Microsoft said demand for AI fueled six percentage points of Azure’s growth, double the amount AI contributed to Azure in the previous quarter. Microsoft now has 53,000 Azure AI customers—one-third of which were new to Azure in the past year, CEO Satya Nadella said on the earnings call on Tuesday.

“It took a decade for Azure to get to $10 billion. At Microsoft AI, it’s at $4.4 billion in one year, so things are happening very fast. I think that it’s year one—I can’t wait to see what year two, year three, and year four bring for Microsoft AI,” Brent Bracelin, Piper Sandler Equity Research Analyst on Cloud, told Yahoo Finance. 

The much-anticipated earnings displayed AI’s potential as a major revenue driver for juggernauts like Microsoft. For the companies trying to integrate these AI technologies to serve customers and employees, however, the technology’s impact is far more fraught. Many of these companies are staring down a line of tough decisions about new security risks, what the technology means for their workforces, how to actually implement the technology, and how to reimagine their products and overall businesses in the name of AI.  

As part of a Canva-commissioned study published yesterday, Harris Poll surveyed more than 1,300 chief information officers from across nine continents and found that 94% plan to increase spending on AI tools in 2024, some by more than 50%. At the same time, 84% said there are already too many AI tools available, making it difficult to figure out what software products they should use, with some companies deploying multiple AI tools with redundant capabilities. The vast majority of CIOs also expressed concerns about not being prepared to train employees on these tools, the increasing complexity of their tech stacks, and opening up new security risks. The results mirror another survey we reported on recently, in which a whopping 66% of C-suite executives across 50 markets said they’re ambivalent about or outright dissatisfied with their organization’s progress on AI and generative AI so far. Many reported having an unclear AI roadmap, lack of necessary talent and skills, and an absence of strategy around how to implement AI responsibly. 

When generative AI left the station, it did so like a bullet train. Thought leaders told business executives to “adopt AI or get left behind,” and those executives—facing investors and boards asking “What’s our AI strategy?”—found themselves needing to come up with answers quickly, even if they didn’t have them. 

“Everyone has jumped on it without considering what needs to change,” Kyle Daigle, chief operating officer of Microsoft-owned GitHub told Eye on AI. 

Working directly with companies to integrate GitHub’s AI-powered programming tool, Copilot, Daigle has seen first-hand how businesses, driven by FOMO, are jumping head-first into AI because it’s the thing to do, not necessarily because they have a strong plan.

“The Fortune 500 and others are trying to come up with their plan on how to use AI,” he said, adding that while they’re seeing customers who are eager to use Copilot, the question is, “Are they choosing to roll it out? Or are they coming up with a plan for a plan?’”

Daigle said the companies finding the most success implementing Copilot are those that are integrating it into their workflows and not just adding an AI button or chatbot window because it’s the hot thing to do. They’re scrutinizing their products, considering their user behavior, and strategizing from the ground up the role AI can play in helping the business achieve its mission.

“I think that there’s kind of this idea that if we wait for the next big AI moment, if we wait for the next AI model, it’s gonna save us. The next leap is the one that’s going to make everything better,” he said. “And I think what’s been proven over the past two years is that’s probably not true. How we use these tools, how we integrate them into our apps and workflows—that’s what makes the big difference.”

And with that, here’s the rest of today’s AI news. 

AI IN THE NEWS

U.S. lawmakers introduce bill targeting nonconsensual porn following Taylor Swift controversy. The DEFIANCE Act (Disrupt Explicit Forged Images and Non-Consensual Edits), was proposed by a group of bipartisan senators and would let people sue over faked pornographic images of themselves. This would include not only AI-created deepfakes but also materials created or edited by any technological means, as well as real pictures that have been modified to look sexually explicit, seemingly applying to older tools like Photoshop as well as AI, according to The Verge. At the state level, a Missouri lawmaker also filed a similar bill named “The Taylor Swift Act,” continuing the pattern of patchwork state regulatory efforts on major AI-related issues. While nonconsensual porn, including the use of deepfakes, has targeted everyday women (even high schoolers) for a while now, the singer’s massive influence and celebrity—and the extent of the harassment against her—has brought an influx of attention that seems to be moving the needle. 

The FCC is moving quickly to criminalize AI-generated robocalls after fake Biden call to New Hampshire voters. Like the bill targeting nonconsensual porn, this action also follows a high-profile incident. Some New Hampshire voters received a fraudulent AI-generated call posing as President Joe Biden and telling them not to vote in the primary election. The proposal would outlaw these robocalls under the existing Telephone Consumer Protection Act, or TCPA, according to NBC. The commission is expected to vote on and pass the action in the coming weeks. 

Italian regulators notify OpenAI that ChatGPT violates EU data privacy rules. The watchdog began investigating OpenAI last year when it temporarily banned ChatGPT in the country for data privacy concerns. Italian regulators have now concluded that ChatGPT violates GDPR, the EU’s General Data Protection Regulation, and has given the company 30 days to respond, according to the Associated Press.

A Microsoft engineer says OpenAI and Microsoft ignored his findings about vulnerabilities in DALL-E 3. OpenAI and Microsoft are engaged in a back-and-forth with a principal Microsoft engineer (and GeekWire) who’s blowing the whistle after he says the companies dismissed and ignored his reports stating he discovered vulnerabilities in DALL-E 3 that allow users to bypass safety guardrails to create violent and explicit images. He also alleges Microsoft lawyers demanded he remove an open letter he posted on LinkedIn to draw attention to the matter. The engineer has now escalated his case to Washington state’s attorney general and Congressional representatives via a letter, also tying the issue to the recent emergence of explicit deepfake images of Taylor Swift. 

AI will worsen the climate crisis, Brooking Institute warns in new report. It will also widen gaps that leave marginalized people disproportionately vulnerable to climate effects, namely low-income and communities of color, the report said. “[Large language models] require ballooning computational resources, just as the energy and infrastructure needs of AI systems are growing themselves thanks to advancements such as generative AI and mounting industry competition.” Those needs include massive amounts of energy, land, water, and expansive supply chains. And there are growing greenhouse gas emissions at every step of the process, from mining materials that go into AI computer chips to making those chips to supplying the power that runs those chips in data centers and keeps the data centers from overheating.

FORTUNE ON AI

Businesses are scrambling to appoint AI leaders—and CAIO compensation packages average well above $1 million —Orianna Rosa Royle

AI is reshaping financial forecasts and disclosure–and making language more important than ever —Richard Torrenzano

Leaders want strategic and critical thinking more than anything. Most of their workers don’t have it —Jane Thier

The fight against greenwashing starts with AI. Here’s why —Anirban Bose

DataSnipper, startup that uses AI to eliminate some of the ‘dread’ in accounting, is valued at $1 billion in latest funding round —by Jeremy Kahn

AI Calendar

Feb. 8-10: World AI Cannes Festival in Cannes, France

Feb. 13-14: MENA Conversational AI Summit in Dubai

Feb. 21: Nvidia reports earnings 

March 18 - 21: Nvidia GTC AI conference in San Jose, Calif.

April 15-16: Fortune Brainstorm AI London

June 25 - 27: 2024 IEEE Conference on Artificial Intelligence in Singapore

BRAIN FOOD

Circling back on Circle to Search. Is Circle to Search just a commerce boost for Google? That’s the question I and many others posed upon the unveiling of the feature for Samsung’s new AI-powered Galaxy smartphones in late January. Now after playing around with the feature on the Samsung Galaxy S24 Ultra, I’m circling back to say it sure seems so. 

For anyone who needs a reminder, Circle to Search is a feature on Samsung’s new Galaxy smartphones (and now the Google Pixel 8) that taps AI to let users search Google about anything from within any app by simply circling it on their screen. Results are then overlaid directly on the current app—no typing or switching apps required. This could theoretically lead users to anywhere on the internet, but I think Google tipped its hat during the on-stage demo, in which a Google VP showed how this could be used to search for a clothing item that catches your eye in an Instagram post. 

In my trials with Circle to Search, I tried coming at it from every angle to see what results Google would turn up. In almost every scenario, I was presented with Google shopping listings to buy the product or one like it. This made sense when I took a picture of and then circled the black boots I was wearing, but it surprised me in many other scenarios. For example, I took a picture of my tote bag, which features a distinctive logo and name for a music organization. When I circled it, Google didn’t search for or offer any information about the music organization. Rather, Google surfaced shopping listings for a bunch of other random tote bags. In another experiment starting with a YouTube video about basketball legends, I circled Magic Johnson as he appeared in the frame, capturing his face and the upper part of his jersey. Google showed me shopping and eBay listings for basketball jerseys rather than identifying the player or providing information about him. 

It’s a reminder that as much as AI is a way to create new user experiences, it’s also just the latest way for companies to sell us stuff. And it mirrors what’s happened to Google search over the years, where the search engine has become more a maze of ads and seller-gamed SEO than a directory of information. Just this week, Alphabet’s stock went tumbling after its quarterly earnings showed the company fell short of analysts' expectations on ad revenue, which has long been the heart of its business. A new channel for promoting shopping and ads, wrapped with flashy AI, makes a lot of sense. 

Personally, I seek out information for knowledge’s sake a whole lot more often than I search for something to buy. If Circle to Search would actually surface relevant and useful information beyond commerce, this could be a new and interesting way of interacting with smartphones and search. Knowing how and where to buy something is actually one of the most intuitive and easy experiences on the internet these days. Finding good information, on the other hand, feels harder and harder every day.