No joke: Google’s AI is smart enough to understand your mood

Amid a flood of new hardware, including the Pixel 7, Pixel Buds Pro, and a new Pixel Tablet, Google unveiled a development at its I/O developer conference that went under the radar: Its AI can now understand jokes.

Jokes, sarcasm, and humor require understanding the subtleties of human language and behavior. For a comedian to say something sarcastic or controversial, the audience can usually discern the tone and know that it is more of an exaggeration, something that has been learned from years of human interaction.

But PaLM, or Pathways Language Model, learned it without being explicitly trained in humor and the logic of jokes. After receiving two jokes, he was able to interpret them and spit out an explanation. In a blog post, Google shows how PaLM understands a novel joke not found on the Internet.


Example of Google’s PaLM learning model that understands and explains a novel joke.


Understanding dad jokes isn’t the ultimate goal of Alphabet, Google’s parent company. The ability to analyze the nuances of natural language and queries means that Google can get answers to complex questions faster and more accurately in more languages ​​and people. This, in turn, can break down barriers and move humans away from communicating with machines through predetermined means and instead interacting more seamlessly. This may include answering questions in one language when finding information in another, or writing code in a program while a person speaks into the model with a specific task.

The power of this technology seemed lost to some in the Twitter space. When searching for #GoogleIO2022, the best results are focused on Pixel hardware. Really, to make people understand the power of its AI technology, Google plans to do it with bespoke chips housed in custom Pixel devices.

“How do you tell my mom, LaMDA or PaLM, let alone what natural language processing is?” said Tuong Nguyen, senior principal analyst at Gartner, a technology research and consulting firm. “The devices serve, in part, as a delivery mechanism for all the amazing things they’re doing.”

Google can combine that with practical examples of how PaLM can make minor annoyances go away. During his appearance on stage at Google I/O on Wednesday, Alphabet CEO Sundar Pichai noted the frustration people who speak lesser-known languages ​​face when trying to find answers to problems in their language. The answer is likely to be online, but it will probably be in English or Spanish.

Pichai showed an example where PaLM was asked in Bengali what pizza toppings are popular in New York. The model can find the answer in English and translate it to the user in Bengali.

“One day, we hope to be able to answer questions on more topics in any language you speak, making knowledge even more accessible on Search and across Google,” Pichai said.

Outperforming Star Trek data?

PaLM is Google’s largest AI model to date and is trained on 540 billion parameters. You can generate code from text, answer a math problem, and explain a joke. It does this through thought chain prompts, which can describe multi-step problems as a series of intermediate steps.

On stage, Pichai described him as a teacher giving a step-by-step example to help a student understand how to solve a problem.

If what Pichai said on stage is accurate, Google has essentially outgrown Star Trek and 400 years of fictional AI development, as evidenced by the character Data, who never really understood the niceties of humor. Furthermore, it appears that Google has caught up with TARS from the movie Interstellar, which takes place in the year 2090, an AI that was so adept at humor that Matthew McConaughey’s character told it to tune in.

PaLM’s ability to understand humor and make logical inferences is helping Google solve novel challenges that previously would have required someone with specific expertise.

In this sense, Google’s AI is a time machine. AI models can accomplish years of study and research in seconds. Time-machine-like companies that can connect people faster or instantly are dominating global markets and changing lives. The value these companies bring to human interoperability is tremendous.

Google, which has been primarily a search and ad sales company, can use its technology to ensure that people continue to turn to Google for answers to their questions. This, of course, will give Google more data points to target advertising to customers. In addition, companies can use their servers and AI to solve complex problems. At I/O, Pichai announced a $9.5 billion publicly available data center and machine learning center in Mayes County, Oklahoma. There, Google Cloud customers can use the facility’s nine exaflops of computing power to run complex models and help solve problems in medicine, logistics, sustainability and more.

On the consumer front, Google has invested in custom development with its Tensor line of chips for its Pixel phones.

“Google is fundamentally a data, AI and search company. And they monetize their attention and their data,” said Avi Greengart, president and principal analyst at Techsponential. “The hardware is an extension of the platform and an ecosystem that drives all of that.”

For Google to fulfill its vision of “ambient computing,” or a future in which people can use computers so intuitively they don’t realize they’re using them, it had to invest in its own hardware.

“Another company might not talk about something coming in the fall, when it’s not even spring, because you’re afraid you’re going to get burned,” Greengart said. “Google is not afraid of that. They prefer to develop this vision that they have interoperable software that does great things.”

Leave a Comment