Artificial intelligence has solved biology’s impossible problems, and AI-powered robots have taught themselves to play soccer.
Machines never get tired. they never go hungry. They learn, grow, and develop superhuman abilities in narrow ways. Most AI systems today do one or two things well. For example, a soccer robot can’t make grocery lists, book trips, or drive a car. The ultimate goal is something called artificial general intelligence. In other words, it is a learning machine that can acquire a wide range of talents.
Some of these talents can be surprisingly human, 60 Minutes correspondent Scott Pelley learned when visiting Google’s new campus in Mountain View, California. Bard, Google’s AI chatbot, It seems to have the total amount of human knowledge. With a microchip that’s more than 100 times faster than his human brain, Bard created a remarkably human story in his five seconds when prompted to “sell.” baby shoes. Never worn.
Byrd’s story featured a man and a stranger whose wife was unable to conceive, grieving after a miscarriage and longing for closure.
“She knew her baby’s soul was always alive,” Byrd wrote when asked to share her story in verse.
Byrd spent months reading nearly everything on the Internet to create a model of what languages would look like, according to Google senior vice president James Manika.
The bard does not recognize himself. AI predicts the most likely words based on all learned words. Still, it doesn’t look like it when Bard explains why it helps people.
“Because it makes me happy,” Byrd said.
60 minutes
Manika explained that artificial intelligence has learned from people, resulting in the emergence of senses and consciousness.
“We are sentient beings. We have feelings, emotions, ideas, thoughts and perspectives. “So as they learn from it, they build patterns from it. No one. These are not sentient beings.”
Like the humans it learned from, the bard is flawed. In an essay written by AI on economics, AI referenced five of his books. Each one was forged. This very human trait, self-confident error, is called hallucinations in the industry. To cure hallucinations, Bard has a “Google it” button that leads to an old-fashioned search. Google also built a safety filter into Bard to screen for hate speech, bigotry, and more.
Google releases a more advanced version of Bard that can connect to its own reasoning, planning, and internet searches so it can do more testing, get more user feedback, and develop a more robust safety layer. is on the way, said Google CEO Sundar Pichai. He walks a narrow gamut on how quickly AI advances will be released.
Critics Rush to AI too fast, but the competitive pressures between tech giants like Google and smaller start-ups are driving humanity into the future, ready or not. Society will need to adapt quickly, whether it’s regulating AI in the economy, laws punishing its abuse, or treaties between nations to make AI safe around the world, Pichai said.
60 minutes
“One of our ways of thinking is how to develop AI systems that are in line with human values. This includes morality. That’s why I think we need to include scientists: ethicists, philosophers, etc.,” Pichai said. “And I think we have to be very thoughtful. And I think these are all things that society needs to understand as we go. It’s not something the company decides.”
The artificial intelligence revolution has been called upon by those who hope it will save humanity. predict fateGoogle is somewhere in the optimistic middle, phasing in AI so civilization gets used to it.
Demis Hassabis, CEO of DeepMind Technologies, has been working with AI for decades and considers it the most important invention mankind has ever made.hassabis sales From DeepMind to Google in 2014One of the reasons for the sale was to access Google’s immense computing power. Brute force computing can very roughly approximate the capabilities of neural networks and the brain.
“We knew all about how the brain works—memory, imagination, planning, reinforcement learning—and we wanted to replicate some of that in an AI system,” says Hassabis.
Harnessing that power, DeepMind created an AI program to predict the 3D structure of proteins. Hassabis says the entire PHD is needed for the average scientist to find his 3D structure of a single protein. DeepMind can predict structures much more quickly.
“In fact, last year we looked at all 200 million proteins known to science,” he said.
Hassabis said DeepMind released the protein database as a “gift to humanity.” It is used in the development of vaccines and antibiotics. The database is also being used to develop new enzymes that eat plastic waste.
AI can make use of all information in the world. Perry wondered if humankind would decline due to the immense power of artificial intelligence.
Manika sees this moment as an inflection point. In a way, he sees his AI being more human and able to answer deeper questions.
“Who are we? What do we care about? What are we good at? How do we relate to each other?” he said. “These are going to be very important issues, and in some cases they will always be. [or] Excited, but probably also anxious. “
