Connect with us

Published

on

Sam Altman, CEO of OpenAI, at the Hope Global Forums annual meeting in Atlanta on Dec. 11, 2023.

Dustin Chambers | Bloomberg | Getty Images

OpenAI CEO Sam Altman has admitted that he was surprised by the popularity of ChatGPT, which was released as a research project a little more than a year ago. His team had spent an entire meeting debating whether it was worth even opening up the chatbot to the public.

As it turns out, OpenAI’s decision to launch ChatGPT in November 2022 became the defining moment for generative artificial intelligence and set the stage for a rush of investments and a mountain of new products and services in 2023.

Some form of generative AI has made its way into virtually every industry, from financial services to biomedical research. Some 95% of utility and energy companies are discussing using generative AI algorithms, according to a July survey

To see the financial effect of the generative AI rush, you need only to look at Nvidia’s bottom line. The chipmaker’s graphics processing units, or GPUs, are at the heart of the large language models created by OpenAI as well as those from Alphabet, Meta and a growing crop of heavily funded startups all battling for a slice of the generative AI pie.

Through the first three quarters of 2023, Nvidia generated $17.5 billion in net income, up more than sixfold from a year earlier. Its stock price jumped 237% this year, far exceeding any other member of the S&P 500.

Generative AI quickly became the buzzy phrase for corporate earnings calls, as every company needed a narrative. Sometimes, the story was painful, such as when digital education company Chegg said in May that it was seeing a “significant spike in student interest in ChatGPT,” which appears to be “having an impact on our new customer growth rate.” The stock plunged 48% in one day following the warning.

Perhaps no company was less prepared for the generative AI boom than OpenAI itself. In November, Altman was suddenly ousted by the board for a dispute that reportedly had to do with his aggressive push to develop new commercial products at the expense of safety. However, Altman quickly returned to the helm after employees threatened to flee and large investors banded together to fight for his reinstatement.

The public spat included a reshuffling of OpenAI’s board and shined a bright light on the debate raging between AI skeptics and evangelists. While advances in generative AI showcase the potential for technology to unlock all sorts of business opportunities and efficiencies, fears of the algorithms’ perceived power gained equal resonance. Some of the real-life harms for minorities and vulnerable populations showed up in FTC proposals, wrongful arrests, contaminated datasets and more.

Here were some of the key areas for generative AI advancements in 2023: 

The Anthropic website on a laptop arranged in New York on Aug. 15, 2023.

Gabby Jones | Bloomberg | Getty Images

Chatbots

ChatGPT opened up the floodgates for investments in chatbots, as it became clear how the input of a few words could produce more thorough and creative responses than ever before.

Nearly two months after launch, ChatGPT broke records as the fastest-growing consumer app in history — until Meta’s Threads dethroned it last summer. ChatGPT now has about 100 million weekly active users, along with more than 92% of Fortune 500 companies using the platform, according to OpenAI. 

Earlier this year, Microsoft plunged an additional $10 billion in the company, making it the biggest AI investment of the year, according to PitchBook, and OpenAI is in talks to sell employee shares at a price that would suggest a valuation of $86 billion. According to Bloomberg, the company is in discussions to potentially raise capital at a valuation of $100 billion or more.

Google was caught off guard by ChatGPT’s success and responded by accelerating the public release of its Bard chatbot, powered by its LLM called LaMDA, which stands for Language Model for Dialogue Applications.

Google has been rolling out new Bard features, including integrations with Google Search and a YouTube extension, and recently released Gemini, the new and buzzy AI model to power Bard. Gemini’s launch this month involved a marketing blitz and controversy over an edited video promoting the model’s capabilities. 

In addition to its internal investments, Google is one of many big names behind Anthropic, an AI startup that’s currently in talks to raise $750 million at a valuation of $18.4 billion. Founded by former OpenAI research executives, Anthropic is the developer of the chatbot Claude.

In July, Anthropic debuted Claude 2, and said it has the ability to summarize up to about 75,000 words, which could be the length of a book. Users can input large datasets and ask for summaries in the form of a memo, letter or story.

As a category, new generative AI chatbots have been used this year to answer questions about business strategy, as well as to design study guides, offer advice on salary negotiation and spark creative writing prompts. They’ve even assisted in writing wedding vows. 

“It’s probably one of the most influential step function changes in technology that we’ve seen,” Jill Chase, investment partner at CapitalG Ventures, told CNBC. Chase said it’s up there with the dawn of the internet and the shift to mobile. “Things like that just open up people’s imaginations,” she said.

Academics and ethicists have voiced significant concerns about the technology’s tendency to fabricate information and to propagate bias. Still, it has quickly made its way into schools, online travel, the medical industry, digital advertising and beyond. Microsoft and IBM have invested increasing amounts in enterprise AI offerings, including development studios for companies to personalize the use of LLMs. 

There are plenty of detractors.

Publishers, artists, writers and technologists have been pushed to pursue legal action against companies behind popular generative AI tools, out of concern that their creative content is being used as free training data. John Grisham, George R.R. Martin and other prominent authors sued OpenAI in September over alleged copyright infringement.

This photo taken on Jan. 31, 2023, shows an artificial intelligence manga artist, who goes by the name Rootport, wearing gloves to protect his identity, demonstrating how he produces AI manga during an interview with AFP in Tokyo.

Richard A. Brooks | Afp | Getty Images

Image and video generation 

Generative AI for images and video emerged in 2022, due to powerful image generators such as OpenAI’s DALL-E 2, Stable Diffusion and Midjourney, and video-generation AI tools from Meta, Google and Amazon

While interest in those technologies continues, progress has waned compared to chatbots, according to Brendan Burke, an analyst at PitchBook. 

“Multimedia content generation has fallen behind language in the pace of progress,” Burke told CNBC. “The initial excitement with Stable Diffusion in 2022 exposed both the general interest but also the drawbacks of AI content generation. Progress has been incremental this year, yet still disappointing for the most sophisticated content creators.”

Meta’s Instagram recently debuted a feature that allows users to change the background of Stories posts using AI. Google and Amazon have incorporated generative AI tools into advertising technology to create more appealing marketing images.

Some industry leaders say the future of generative AI is “multimodal,” bringing the various mediums together.

“The world is multimodal,” Brad Lightcap, OpenAI’s operating chief, told CNBC in a recent interview. “If you think about the way we as humans process the world and engage with the world, we see things, we hear things, we say things. The world is much bigger than text. So to us, it always felt incomplete for text and code to be the single modalities, the single interfaces that we could have to how powerful these models are and what they can do.”

Agents and assistants

After the chatbot comes the agent.

It’s not just about getting sophisticated answers, but it’s also about using generative AI to be productive in completing tasks. That could be scheduling a group hangout by scanning everyone’s calendar to make sure there are no conflicts, booking travel and activities, buying presents for loved ones or doing a specific job function such as outbound sales.

Last month, OpenAI announced custom GPTs, or customized, niche versions of ChatGPT that users can personalize for getting travel recommendations, recipe help or startup advice. However, the company chose to delay the release of the platform that would popularize different use cases — the “GPT store” — until next year. 

One type of AI assistant that has gained popularity is for coding. Take, for example, Microsoft’s GitHub coding repository. GitHub CEO Thomas Dohmke wrote in a blog post earlier this year that an average of 46% of all code on GitHub, “across all programming languages,” was AI generated.

Last month, GitHub introduced a more expensive version of its Copilot assistant that can explain and provide recommendations about internal source code.

“Copilot, when it started at the very beginning, was thought to be a tool that could help developers write docs,” Kyle Daigle, GitHub’s chief operating officer, told CNBC in an interview. In the past year, he said, the company has expanded the technology, looking for more places “to help developers collaborate and work together and solve problems outside of just the code.” 

But PitchBook’s Burke said coding assistants are in their very early days and currently can only do “a small part” of a developer’s work. That’s true in the broader world, he said. 

“Users have found how little AI can do for them this year,” he said. “AI knows a lot, but it can’t do a lot yet. We’re still far away from AI truly being able to do the complex tasks that people are used to doing in their personal lives and at work. That has been shown by the struggles of AI agents this year.” 

Nvidia CEO Jensen Huang speaks at the Supermicro keynote presentation during the Computex conference in Taipei on June 1, 2023.

Walid Berrazeg | Sopa Images | Lightrocket | Getty Images

Looking ahead 

Overall, 2023 was a big year for consumer excitement surrounding generative AI and for adoption of a few popular products. But business success stories have been few and far between.

“It was an especially transformative year from a consumer perspective where AI became much more tangible than before,” Grace Isford, a partner at Lux Capital, said in an interview. “AI is nothing new, but the awareness — and in turn, the adoption — has skyrocketed. Many more hackers and builders are leveraging the technology and the really exciting advancements into products.” 

CapitalG’s Chase said the consumer fascination with the space has allowed people to “see what was possible” in AI, allowing for a “cake tasting” of sorts and a teasing of the imagination.

Early in the year, people “extrapolated out early exploration of that technology into lasting and enduring use cases,” Chase said. She added that there hasn’t been a straight line from early adoption and widespread use of one or two products to mainstream popularity. Companies and developers are now going back to doing research and development to “build the right infrastructure and tooling” that can hopefully lead to mass adoption.

“I think that will happen over the next year,” she said. “I think some people thought it would happen this year.”

In 2023, it’s clear that the overwhelming beneficiary from all the hype was Nvidia. The challenge for the coming year and beyond is for businesses to show that their hefty spending on those advanced GPUs and the models they power can lead to the development of products that allow more companies to share in the wealth.

“I thought that the excitement at the end of last year would quickly translate into enterprise adoption, but the reality is that very few companies have launched generative AI applications into production and experiments aren’t quickly translating into reliable applications,” Burke said. “We’re still looking at an outlook where companies may not widely deploy products until later next year or even the following year.”

WATCH: There’s good news for Microsoft about corporate AI spending plans

There's good news for Microsoft about corporate AI spending plans in 2024


Don’t miss these stories from CNBC PRO:

Continue Reading

Technology

Britain expands AI safety institute to San Francisco amid scrutiny over regulatory shortcomings

Published

on

By

Britain expands AI safety institute to San Francisco amid scrutiny over regulatory shortcomings

An aerial view of the city of San Francisco skyline and the Golden Gate Bridge in California, October 28, 2021.

Carlos Barria | Reuters

LONDON — The British government is expanding its facility for testing “frontier” artificial intelligence models to the United States, in a bid to further its image as a top global player tackling the risks of the tech and to increase cooperation with the U.S. as governments around the world jostle for AI leadership.

The government on Monday announced it would open a U.S. counterpart to its AI safety summit, a state-backed body focused on testing advanced AI systems to ensure they’re safe, in San Francisco this summer.

The U.S. iteration of the AI Safety Institute will aim to recruit a team of technical staff headed up by a research director. In London, the institute currently has a team of 30. It is chaired by Ian Hogarth, a prominent British tech entrepreneur who founded the music concert discovery site Songkick.

In a statement, U.K. Technology Minister Michelle Donelan said the AI Safety Summit’s U.S. rollout “represents British leadership in AI in action.”

“It is a pivotal moment in the U.K.’s ability to study both the risks and potential of AI from a global lens, strengthening our partnership with the U.S. and paving the way for other countries to tap into our expertise as we continue to lead the world on AI safety.”

The expansion “will allow the U.K. to tap into the wealth of tech talent available in the Bay Area, engage with the world’s largest AI labs headquartered in both London and San Francisco, and cement relationships with the United States to advance AI safety for the public interest,” the government said.

San Francisco is the home of OpenAI, the Microsoft-backed company behind viral AI chatbot ChatGPT.

The AI Safety Institute was established in November 2023 during the AI Safety Summit, a global event held in England’s Bletchley Park, the home of World War II code breakers, that sought to boost cross-border cooperation on AI safety.

The expansion of the AI Safety Institute to the U.S. comes on the eve of the AI Seoul Summit in South Korea, which was first proposed at the U.K. summit in Bletchley Park last year. The Seoul summit will take place across Tuesday and Wednesday.

The government said that, since the AI Safety Institute was established in November, it’s made progress in evaluating frontier AI models from some of the industry’s leading players.

It said Monday that several AI models completed cybersecurity challenges but struggle to complete more advanced challenges, while several models demonstrated PhD-level knowledge of chemistry and biology.

Meanwhile, all models tested by the institute remained highly vulnerable to “jailbreaks,” where users trick them into producing responses they’re not permitted to under their content guidelines, while some would produce harmful outputs even without attempts to circumvent safeguards.

Tested models were also unable to complete more complex, time-consuming tasks without humans there to oversee them, according to the government.

It didn’t name the AI models that were tested. The government previously got OpenAI, DeepMind, and Anthropic to agree to opening their coveted AI models up to the government to help inform research into the risks associated with their systems.

The development comes as Britain has faced criticism for not introducing formal regulations for AI, while other jurisdictions, like the European Union, race ahead with AI-tailored laws.

The EU’s landmark AI Act, which is the first major legislation for AI of its kind, is expected to become a blueprint for global AI regulations once it is approved by all EU member states and enters into force.

Continue Reading

Technology

$9 billion travel tech firm Navan on track to hit profitability this year and ‘not far’ from IPO, CEO says

Published

on

By

 billion travel tech firm Navan on track to hit profitability this year and 'not far' from IPO, CEO says

TripActions CEO Ariel Cohen

TripActions

The boss of travel and expense management platform Navan told CNBC he’s preparing the company to get its business into shape for an eventual initial public offering this year, in another sign leaders of privately-held startups are getting more optimistic about their prospects in the public markets.

Asked about when Navan would choose to go public, the firm’s CEO and co-founder Ariel Cohen said the company is close to reaching that milestone. “We can see the signals,” he said, adding that Navan has been adjusting its leadership structure and making changes to its board in a signal of maturity.

Last month saw Navan announce the return of Rich Liu, formerly Navan’s chief revenue officer and “an expert on scaling companies from seed to IPO and beyond,” to the business as CEO of Navan Travel, the company’s travel division.

Amy Butte, the former chief financial officer of the New York Stock Exchange who oversaw the U.S. exchange operator’s transition to a public company in 2006, was also appointed to Navan’s board of directors as audit committee chair.

“I don’t want to give a date,” Cohen told CNBC, adding that he’s not even told his own family a date for when he expects Navan to go public — let alone his board and Navan employees. “At the end of the day, there are things that are out of my control.”

“The market can collapse. There are elections in the U.S. There are wars. So I never actually promise things to people if I don’t know that the delivery is in my control,” he added.

While Cohen wouldn’t commit to a date for Navan’s eventual IPO, he said the business was “not far” from being ready for a stock market listing. The company is on track to become cash-flow positive and achieve profitability for the first time this year, he said.

“The timing will need to include several things,” he said. “Today, in this market, to be public, you need to be profitable. We are not far from that, but we are not there. We’re going to be there this year. And it’s not easy to do it while you’re growing fast.”

Cohen said he’s also keeping a wary eye on the market — but added that although, previously, investors would have seen a company like his as dependent on buoyant market sentiment surrounding technology, today he sees the firm as “mature enough” to go public independent of the market backdrop.

Navan CEO Ariel Cohen talks partnering with Citi

Navan is now growing revenues by around 40% on average, according to Cohen, with the company’s fintech business seeing faster growth (100%) than its travel business (30%).

Founded in 2015 as TripActions, Navan began life as a travel management platform for businesses, seeking to provide a smoother experience to travel agents and incumbent players like American Express, BCD Travel, and SAP Concur. The company counts the likes of Unilever and Christie’s as clients.

The firm subsequently expanded into expensing and payments with solutions for automating linking credit cards to a single platform and automating expenses.

Navan is backed by major investors including Andreessen Horowitz, Coatue, Goldman Sachs, and Lightspeed. Navan has raised more than $1.5 billion in venture funding to date and was last valued at $9.2 billion. It competes with Spanish startup TravelPerk, which was most recently valued at $1.4 billion.

Navan introduced a big evolution of that product last year with the arrival of Navan Connect, a new expensing product.

Most corporate card startups, like Brex and Ramp, offer users their own branded corporate smart cards. But Navan’s Connect feature, which it’s rolled out in partnership with Citi, lets businesses offer automated expense management and reconciliation without having to change corporate card provider.

Like other tech firms, Navan has been making a big investment into artificial intelligence. The company rolled out its own AI personal assistant, called Ava, last year. The tool uses generative AI to help travelers, travel admins, and finance managers make travel plans and budget effectively.

Ava — which stands for automated virtual assistant — now processes around 150,000 monthly chats, more than 35% of which are managed to completion as of April 2024, according to Navan.

Cohen said Navan is planning to roll out an even more personalized version of Ava’s AI assistant, which can generate travel plans for someone based on their past behavior, to even greater accuracy in six months’ time.

Navan was named on the 2024 edition of CNBC’s Disruptor 50 list.

Continue Reading

Technology

Musk launches SpaceX’s Starlink internet services in Indonesia, says more investments could come

Published

on

By

Musk launches SpaceX’s Starlink internet services in Indonesia, says more investments could come

Tech billionaire Elon Musk (2nd L) speaks next to Indonesia’s Health Minister Budi Gunadi Sadikin (L) during a ceremony held to inaugurate satellite unit Starlink at a community health center in Denpasar on Indonesia’s resort island of Bali on May 19, 2024. Musk launched on May 19 his Starlink service on Indonesia’s resort island of Bali as the country aims to extend internet to its remote areas. Millions of people in Indonesia, a vast archipelago of more than 17,000 islands, are not currently hooked up to reliable internet services. (Photo by SONNY TUMBELAKA / AFP) (Photo by SONNY TUMBELAKA/AFP via Getty Images)

Sonny Tumbelaka | Afp | Getty Images

Elon Musk has launched SpaceX’s Starlink satellite internet services in Indonesia as the Southeast Asian nation seeks to boost internet connectivity in remote areas.

The inauguration took place at a community health center in Bali on Sunday, with the SpaceX founder telling local media he was “very excited” to bring internet services to areas with limited or no connectivity and that internet connectivity can be “a life-saver to remote medical clinics.”

“It is really important to emphasize the importance of internet connectivity and how much of a life-changer and a life-saver it can be,” Musk told local media on Sunday.

SpaceX, which manufactures and launches advanced rockets and spacecraft, is also a developer of Starlink satellites which provide internet connectivity to remote locations. Starlink is already available in Southeast Asia in Malaysia and the Philippines.

Indonesia, the world’s largest archipelago with more than 17,000 islands, faces an urban-rural connectivity divide where millions of people living in rural areas have limited or no access to internet services.

Communication and Informatics Minister Budi Arie Setiadi previously said Starlink would help Indonesia extend internet access to regions not covered by local internet providers, according to Indonesian news agency Antara.

Elon Musk meets with China's Premier Li Qiang to discuss Tesla, full-self driving and restrictions

The minister also tried to dispel concerns that Starlink’s entry would hurt local internet providers.

“When you have access to the internet, you can learn anything,” said Musk, who is also the CEO of electric carmaker Tesla. “But if you don’t have internet connectivity, it is very difficult to learn.”

“And if you have goods and services you wish to sell to the world, and even if you are in a remote village, you can now do so with internet connection. So you can bring a lot of prosperity to [remote] communities,” said Musk.

When asked about whether he has plans to invest in Indonesia’s electric vehicle sector as well, Musk said he was focused on Starlink first.

“It’s very likely that my companies will invest in Indonesia,” said Musk.

Indonesia minister Luhut Pandjaitan said Musk is considering to set up an EV battery plant in the country, after the tech leader met with President Joko Widodo on Monday, according to Reuters.

Continue Reading

Trending