This is how Google’s AI is developing, which “only empathy can resist”

You probably remember tricks like using quotes or hyphens for better Google searches, articles full of tricks that were very popular on the Internet not so long ago. The Internet giant went from announcing with great fanfare that its search engine could correct spelling errors to make searching easier, to introducing a new Gemini-based system in less than a decade. This is the revolution of artificial intelligence, which this year will fill the world with virtual assistants capable of solving a thousand and one tasks, but this It also helps combat climate change and treat diseases.

Google has gone from fixing two dancing letters in a word to being able to summarize thousands of emails in seconds or explain the contents of a PDF to save technicians time. He is not alone in this revolution; in the same week, OpenAI introduced ChatGPT4o, not without controversy due to the similarity of the assistant’s voice to Scarlett Johansson.

Ability to analyze and quickly summarize hours-long videos, help your child do his homework with detailed explanations and not only the answer, translating conversations from multiple languages ​​in real time. Today, this technology offers endless assistance tasks, and many people either haven’t heard of it or don’t even know where to start using it.

This crazy feeling caused by the latest advances in artificial intelligence is shared by those inside this new world. “It’s real,” explains Conchita Díaz, head of AI training at Google EMEA, in an interview with EL ESPAÑOL-Omicrono during the Google Cloud celebration in Madrid this week. “We (Google) launched Gemini in December, released an improved version two months later, another version last week, that is, technology development is exponential.” But she is convinced that”It will be something natural for the user because he is already interacting with it.Answer: when you go to Netflix and it recommends other content based on your tastes, that’s already generative AI, or when, let’s end this interview, I’m going to order an Uber and they use generative AI, that’s what we’re doing today already integrated” .

“More hours of sleep”

We will have no choice but to get used to it because the changes seem unstoppable. According to a study presented by Google during this event, if the implementation of artificial intelligence were delayed by five years, the GDP potential in Spain would be reduced from 8% to 2%, and if these new tools were to be adopted wholeheartedly, the effect would be 100,000-120 000 million euros in Spain’s GDP.

This directly impacts jobs. Unlike the fear transmitted so far, The study guarantees that in Spain, 58% of jobs will coexist using generative artificial intelligence. and only 6% of jobs are likely to disappear, although they claim that the new jobs created will compensate for these losses.

The mass disappearance of jobs is one of the biggest concerns society has with artificial intelligence. Leaders like Sam Altman recognized this opportunity. In his opinion, only teachers will be saved, since people prefer to deal with other human mentors.

Conchita Diaz emphasizes in the interview that empathy will be the one thing that machines won’t be able to do, and that it will help humans continue to be essential in countless professions. “Augmented intelligence will help us solve problems that we, as a society, are unable to solve today. solve problems such as climate change, natural disasters, early detection of diseases,” and insists: “Even gaining hours of sleep if we have an assistant who automates tasks that do not require our intelligence or creativity.”

Saving working time

To understand how artificial intelligence can help professionals without interrupting their work, Diaz remembers that it’s not just about checking email or asking a search engine a question. There are different types of artificial intelligence: traditional, because it was developed first and trained on a single task; Generative AI is the most talked about AI right now and trains on more data to have a global view. This allows them to solve a wide range of problems. From there, the move was made to multimodal technologies capable of processing video, images, text and code.

“In many cases, we are talking about traditional AI,” Diaz emphasizes, citing the example of an image-trained algorithm developed for the Lays brand, in which the machine recognizes which potatoes can be put into a bag for consumption and which should be thrown away. Another example of this traditional AI is Global Omnium on the Mediterranean coast. Using images collected by ships from the coastal seabed, the AI ​​is sampling in search of Posidonia and Cymodocea meadows to help the Valencian Department of the Environment protect this endemic Mediterranean species.

On the other hand, there is generative AI, which is what AlphaFold is based on. This biomedical tool was launched in 2020, its second generation came two years later, and continuing in the same vein, AlphaFold 3 was introduced this year. This new version can predict not only proteins, but also… structure of proteins with chemical modifications necessary for their functionalityas well as interactions between proteins and other smaller molecules and, ultimately, nucleic acids.

All this biological complexity is leading to faster drug development. “This means that our work is threatened by artificial intelligence, quite the opposite,” says Carlos Fernandez Tornero, a research scientist at CSIC, during the Google Cloud Summit. “This allowed us to complete in a few months what would have taken several years of effort.and thus be able to devote more time to studying other things, such as molecular dynamics, which artificial intelligence still does not seem to protect.”

CSIC scientists in Spain were able to study the parasite that transmits sleeping sickness, a serious problem in Africa that is deadlier than Ebola and which could spread to other regions due to climate change.

AlphaFold is also used to treat cancer,” he said.macaw “This is one of the most important innovations we are doing with artificial intelligence,” Diaz says. As she herself points out, access to data is critical here, like the fuel to launch an AI rocket. “I hope that with regulation this will happen. people who are experts on this issue so that it is not regulated only with “yes” or “no”, but rather by looking at specific cases to be able to use important information that will allow progress in the fight against cancer.

Study in Spain

So that there are experts regulating AI, so that people do not lose their jobs and do not allow abuses, part Google’s Commitment to Advancing Learning. Thus, they proposed that Spain should have 1.3 million people trained to use AI to their advantage.

To this end, this week they presented a series of training courses for all levels: from the user who wants to start from scratch, through startups who want to integrate it into their work, to professionals who want to dedicate themselves to it from now on.

At the consumer level, it is enough to understand how you can adapt them to your life or business to make your life easier. “I would advise everyone to go a little deeper, Once they get in a little they’ll want to know a little more“, says this expert. Google has launched several certificates: for learning the professional fundamentals of AI, which will have a thousand scholarships for the unemployed or, for example, for those who want to master data analysis, one of the most advanced levels.