OpenAI rolled out ChatGPT in November 2022, amassing over 100 million users. On the other hand, tech gargantuan Google launched its ChatGPT-like Bard to the wider public only this week, removing its waitlist and expanding access to 180 countries and territories.
Despite its seemingly endless resources, manpower, and history of AI innovation, did Google fall behind?
Image credit: Timmy Loen
One of its earliest AI projects was Google Brain, an exploratory lab established in 2011 that focuses on deep learning, computer vision, and natural language processing, among others. The unit eventually collaborated with DeepMind, a British company specializing in artificial general intelligence that Google acquired in 2014 for over US$500 million.
Both units are now under the umbrella of Google AI. AlphaGo, the first computer to beat a human champion at the complex Go game, and the computer vision app DeepDream are among the innovations that came out of Google Brain and DeepMind.
6Interestingly, Google AI also created the conversational bot Meena, whose existence was announced in 2020. Over a year later in May, Google introduced LaMDA, its ChatGPT-like large language model, via a post on its official blog.
However, the company was hesitant to release Meena and LaMDA to the public because of ethical issues and potential reputational damage.
Such fears are not unfounded: These programs have previously been accused of becoming sentient, raising existential questions and indicating that Google might lose control over its creations.
OpenAI, though backed by Microsoft, is a relatively minuscule company that has under 1,000 employees. In comparison, Google employs close to 200,000 people on a full-time basis and its parent firm, the Nasdaq-listed Alphabet, has a market cap of 1.48 trillion dollars. As such, OpenAI has less to lose and is in a better position to take huge risks.
And there are risks a-plenty: Google was warned by its ethicists not to launch the Bard chatbot, for instance. And even though it was aware of the risks, Microsoft laid off its entire AI Ethics and Society team earlier this year.
Faith in the tech has gotten so weak that even neural network researcher and AI pioneer Geoffrey Hinton recently quit Google so he could be free to warn the public of the tech’s dangers.
But for Microsoft, ChatGPT’s public release was a blessing. After being massively out-innovated and losing its search engine dominance to Google, Microsoft now has a first-mover advantage in the AI space.
Because of the hype and revenue generated by ChatGPT, Google may have felt pressured to release its research to the public. That said, the company is taking pains to point out that Bard is still an experiment, with even Google CEO Sundar Pichai admitting that he doesn’t “fully understand” how the tool works.
Earlier in February, Bard also made a mistake when it misidentified a fact about the James Webb Telescope. As a result, Alphabet’s market value took a US$100 billion hit.
But it seems that Bard was worth the wait.
Image credit: Bard
The model uses real-time data as it draws information from Google Search, whereas ChatGPT was trained on a dataset limited to 2021.
Bard’s larger dataset also means that answers are more accurate and specific. Try asking Bard if it knows who you are – chances are, it can identify you if you have any online presence. The same can’t be said for OpenAI’s tool.
With access to this data, an over 90% market incumbency on search engines, cloud-based documents, video streaming via YouTube, and email, Google may be set to dominate the AI chatbot space, potentially eclipsing Microsoft’s gains despite its head start.
Google is also looking to incorporate Bard into its search engine, create image-based features through Lens, and integrate it into other products, such as Docs, Gmail, and Sheets.
Ultimately, this will push other Big Tech firms to innovate even more, engaging in bidding wars for independent AI companies. Case in point: Google invested heavily in US-based AI lab Anthropic earlier this year.
As Big Tech monopolizes the space, more and more risks may be taken, and transparency measures and public interest could be overlooked in favor of revenues.
Perhaps the most frightening prospect is how the AI chatbots, which have a singular, omniscient, and assertive voice, directly source info from the Internet and can further spread misinformation and half-truths. But the question is, who will bell the cat?
This was published as a part of AI Odyssey, a section on generative AI developments featured in Tech in Asia’s emerging tech newsletter.
Delivered every Tuesday via email and through the Tech in Asia website, this free newsletter breaks down the biggest stories and trends in emerging tech. If you’re not a subscriber, get access by registering here.