The late comedian Rodney Dangerfield was known for his punch line, "I don't get no respect." Meta Platforms (META 0.21%) just might be the Rodney Dangerfield of the artificial intelligence (AI) race.
Microsoft (MSFT 0.45%) has been in the spotlight lately with its integration of OpenAI's ChatGPT with Bing. The tech giant's moves have Alphabet (GOOG 1.12%) (GOOGL 1.33%) playing catch-up. However, many investors know not to dismiss Alphabet's chances with its deep AI expertise and its Bard chatbot soon to launch.
Meanwhile, Meta doesn't enter the AI conversation as often these days. But could it be the underdog AI stock that makes investors the most money over the long run?
Image source: Getty Images.
Meta's AI creds
Meta CEO and founder Mark Zuckerberg stated in the company's fourth-quarter conference call earlier this month that AI serves as "the foundation of our discovery engine and our ads business." That's not just spin. Meta has long been a pioneer of AI, incorporating the technology extensively into its products.
The news feeds and ad placements on Facebook and Instagram rely heavily on AI. Facebook was a pioneer in image recognition. Although it eventually stopped using facial recognition to tag people in photos (because of bias and privacy issues), Meta continued to advance image recognition algorithms and computer vision.
Today, Meta remains a leader in AI -- even if that might not seem to be the case with the focus on Microsoft and Google. The company's researchers are focusing on multiple AI fronts, including natural language processing, reinforcement learning (an approach based on rewarding positive behavior and penalizing negative behavior), and speech recognition and synthesis.
Meta's innovations haven't been limited to software. In 2022, the company launched the AI Research SuperCluster, one of the fastest AI supercomputers in the world. Meta's scientists plan to eventually use the system to train AI models "with trillions of parameters." By comparison, Open AI's ChatGPT was trained on 175 billion parameters.
Why it could be the biggest winner
Zuckerberg is making generative AI one of Meta's top priorities this year. He said in Meta's Q4 call that one of his goals for the company is "to become a leader in generative AI in addition to our leading work in recommendation AI." He didn't go into detail on what Meta is doing but did say that nearly all of its products will use the new technologies.
Huge sets of training data are critical for improving AI systems. That could give Meta an advantage. The company's Facebook, Instagram, Messenger, and WhatsApp applications combined are used by nearly 3 billion people on average every day. Few rivals can match the conversational data access that Meta has.
Meta could also have an easier time monetizing generative AI than either Alphabet or Microsoft will. It's not hard to envision the company launching chatbots to help businesses interact with customers on Messenger and WhatsApp or rolling out AI technology that helps creators develop more engaging content for Facebook and Instagram.
Don't overlook Meta's big bet on the metaverse. AI will be important in making Zuckerberg's metaverse vision a reality. It's even possible that Meta's work in developing the metaverse could lead to AI advances that catapult the company past its rivals.
What could go wrong
It's certainly no slam dunk for Meta to win in the AI race. Alphabet, Microsoft, or another company could simply build a better mousetrap than Meta does.
Meta could also make mistakes along the way. For example, the company actually launched its own chatbot called Blenderbot last year before OpenAI released ChatGPT. However, Blenderbot was lambasted by users for its poor responses. A few more blunders like Blenderbot could doom Meta to also-ran status.
But many seem to be underestimating Meta's chances in the AI race. That could prove to be short-sighted, especially with the company's AI expertise and continued investments. Meta does appear to be an underdog for now. That won't necessarily be the case in five or 10 years, though.
