Artificial intelligence solved an impossible problem in biology and robots powered by AI taught themselves to play soccer.
Exploring the human-like side of artificial intelligence at Google | 60 Minutes
The machines never get tired. They never get hungry. They learn, and grow, developing superhuman abilities in narrow ways. Most AI systems today do one or maybe two things well. The soccer robots, for example, can’t write up a grocery list or book your travel or drive your car. The ultimate goal is what’s called artificial general intelligence: a learning machine that can score on a wide range of talents.
© Provided by CBS News
Some of those talents can seem shockingly human, 60 Minutes correspondent Scott Pelley learned during a visit to Google’s new campus in Mountain View, California. Bard, Google’s AI chatbot, appears to possess the sum of human knowledge. With microchips more than 100 times faster than the human brain, Bard took 5 seconds to create a deeply human tale when given the prompt: For sale. Baby shoes. Never worn.
Bard’s story featured a man whose wife could not conceive and a stranger, grieving after a miscarriage and longing for closure.
“She knew her baby’s soul would always be alive,” Bard wrote when asked to share the story in verse.
Over the course of several months, Bard read almost everything on the internet and created a model of what language looks like, Google Senior Vice President James Manyika said.
Bard isn’t aware of itself; the AI predicts the most probable words based on everything it’s learned. Still, it doesn’t seem that way when Bard explains why it helps people.
“Because it makes me happy,” Bard said.
© Provided by CBS News Bard is demonstrated at Google / Credit: 60 Minutes
The appearance of sentience and awareness comes because artificial intelligence has learned from people, Manyika explained.
“We’re sentient beings. We have beings that have feelings, emotions, ideas, thoughts, perspectives. We’ve reflected all that in books, in novels, in fiction,” Manyika said. “So, when they learn from that, they build patterns from that. So, it’s no surprise to me that the exhibited behavior sometimes looks like maybe there’s somebody behind it. There’s nobody there. These are not sentient beings.”
Like the humans it’s learned from, Bard is flawed. In an essay the AI wrote about economics, it referenced five books; each one was fabricated. This very human trait, error with confidence, is called, in the industry, hallucination. To help cure hallucinations, Bard features a “Google it” button that leads to old-fashioned search. Google has also built safety filters into Bard to screen for things like hate speech and bias.
Google is holding back on releasing more advanced versions of Bard that can reason, plan and connect to internet search on their own so that the company can do more testing, get more user feedback and develop more robust safety layers, Google CEO Sundar Pichai said. He’s walking a narrow line in how quickly AI advancements are released.
Critics argue the rush to AI comes too fast, but competitive pressure, among tech giants like Google and smaller start ups, is propelling humanity into the future — ready or not. Society needs to adapt quickly, with regulations for AI in the economy, laws that punish abuse and treaties between nations to make AI safe in the world, Pichai said.
© Provided by CBS News Scott Pelley with Google CEO Sundar Pichai / Credit: 60 Minutes
“You know, one way we think about: How do you develop AI systems that are aligned to human values– and including– morality? This is why I think the development of this needs to include not just engineers, but social scientists, ethicists, philosophers and so on,” Pichai said. “And I think we have to be very thoughtful. And I think these are all things society needs to figure out as we move along. It’s not for a company to decide.”
The revolution in artificial intelligence is at the center of a debate ranging from those who hope it will save humanity to those who predict doom. Google lies somewhere in the optimistic middle, introducing AI in steps so civilization can get used to it.
Demis Hassabis, CEO of DeepMind Technologies, has spent decades working on AI and views it as the most important invention humanity will ever make. Hassabis sold DeepMind to Google in 2014. Part of the reason for the sale was to gain access to Google’s immense computing power. Brute force computing can very loosely approximate the neural networks and talents of the brain.
“Things like memory, imagination, planning, reinforcement learning, these are all things that are known about how the brain does it, and we wanted to replicate some of that in our AI systems,” Hassabis said.
With that power, DeepMind created an AI program to predict 3D structures of proteins. It takes an average scientist their entire PHD to find the 3D structure of a single protein, Hassabis said. DeepMind can predict the structures much more quickly.
“And actually, over the last year, we did all the 200 million proteins that are known to science,” he said.
DeepMind made its protein database public as a “gift to humanity,” Hassabis said. It’s been used in developing vaccines and antibiotics. The database has also been used for developing new enzymes to eat plastic waste.
AI can utilize all the information in the world. It left Pelley wondering if humanity is diminished by the enormous capability of artificial intelligence.
Manyika views this moment as an inflection point. In some ways, he thinks AI raises humanity to answer deeper questions.
“Who are we? What do we value? What are we good at? How do we relate with each other?” he said. “Those become very, very important questions that are constantly gonna be, in one case [or] sense exciting, but perhaps unsettling too.”