Chevron icon It indicates an expandable section or menu, or sometimes previous / next navigation options. HOMEPAGE

IBM's CEO says it's leaving consumer AI like facial recognition to other companies, as it doubles down on pitching Watson to businesses

IBM CEO Arvind Krishna
CEO Arvind Krishna said IBM is leaving consumer AI applications to other companies and zeroing in on businesses. Brian Ach/Getty Images for Wired

  • CEO Arvind Krishna said IBM is not building consumer-oriented AI tools like facial recognition.
  • Instead, IBM is placing its bets on AI for the enterprise, helping businesses analyze their data.
  • The strategy is a departure from IBM Watson's early days playing "Jeopardy!" and creating recipes.
Advertisement

Arvind Krishna, the CEO of IBM who just reached his one-year anniversary leading the company, said Monday that Big Blue is not building consumer use cases for its flagship Watson artificial intelligence platform, and instead solely focused on business applications.

"We are not going to do consumer AI," he said. "I think that the consumer companies are doing great at consumer AI, whether it's photograph recognition, whether it's recognizing a face inside somebody's photograph, whether it's speech-to-text and so on."

IBM's sole focus is "enterprise AI," Krishna continued, saying that Watson will be "learning from small amounts of data" and utilizing natural language processing to answer specific business questions for customers.

While Krishna didn't name names, companies like Google, Facebook, and Apple have all invested heavily in AI, while the reference to speech-to-text may be a reference to Microsoft's planned $19.7 billion acquisition of Nuance. Facial recognition, in particular, has proven to be a controversial technology, with critics charging that its use in law enforcement can be abused and weaponized against innocent civilians.

Advertisement

"We have been very clear that AI should not be used for certain things," Krishna said of facial recognition technology, which IBM stopped developing and selling last year.

His comments on IBM's AI plans were to reporters and analysts earlier this week, where he also said that IBM's return to the office could stretch on as long as 18 months from now and play out differently in other countries.

On Tuesday, in his second keynote as CEO during IBM's annual Think conference, Krishna underscored the company's big bets on hybrid cloud and AI, especially as it aims to reinvent itself in a technology landscape dominated by cloud heavyweights Amazon, Microsoft, and Google. IBM's focus on hybrid cloud, kicked off by its purchase of open source software company Red Hat in 2019, will be a major theme for the company in the years ahead.

In laying out his vision for IBM, Krishna is also articulating a more refined vision for Watson, which has at various points been hyped as a "Jeopardy!"-playing, recipe-creating, music-composing super-powered AI platform. But that didn't always translate to business success: IBM shut down its AI tool for drug discovery in 2019 after sluggish sales, and reportedly encountered hesitancy from banks when trying to sell into the financial sector.

Advertisement

Now, Krishna is pitching a more toned-down vision for Watson: It's a useful tool for asking specific kinds of business questions, and getting useful answers — not the kind of general-purpose AI found in science fiction.

"You might be answering a few hundred kinds of questions. You're not looking at trying to answer general purpose Q&A over there," he said. "And that is where our focus is going to be. And that's where we see a lot of benefit for our clients."

IBM enterprise AI growth has 'completely changed' over past year, top exec says

IBM senior vice president Rob Thomas, who oversees the cloud and data platform unit, told Insider in a recent conversation that Watson's business AI focus has three dimensions: Using natural language processing to understand communications like chat and email, building an AI that is trustworthy and free from bias, and automating business functions like customer service and IT operations.

After an early period of "experimentation" in creating a trust framework around Watson two years ago, Thomas says that has "completely changed in the last nine months." 

Advertisement

New customers have recently signed on to Watson, he said, though cited its sensitivity as a reason "not everybody wants to talk publicly yet." Companies are so far using Watson in their hiring processes to remove bias, and banks that are running tens of thousands of AI models "need AI to manage the AI," Thomas said.

Automating IT operations via a philosophy called AIOps is also a critical part of IBM's hybrid cloud strategy, Thomas says, and the main driver behind its recent acquisition of AI firm Turbonomic for over $1.5 billion and application performance monitoring company Instana.

"Turbonomic and Instana can both look at what's happening in the overall environment, to bolster our overall understanding that can feed into an engine like Watson," Krishna said. "These capabilities allow us to manage in a hybrid cloud environment a lot better than we have so far."

And as it continues its reinvention process, IBM is also betting that Watson's AI capabilities will become the next big thing in the IT market.

Advertisement

"When I think about artificial intelligence, I think the same way as maybe there was electrification at the turn of the last century," Krishna said. "I think we can say AI is going to be what happens and gets infused into every business this century."

IBM AI Cloud Computing
Advertisement
Close icon Two crossed lines that form an 'X'. It indicates a way to close an interaction, or dismiss a notification.

Jump to

  1. Main content
  2. Search
  3. Account