The question comes back constantly, and with increasing insistence starting from the last months of 2022, every time we write (or speak) of ChatGPT, Stable Diffusion or any other generative AI. She comes back despite the fact that the answer has already been given several times: “But when does Google come out of her?”. Written like this, in social Italian, below the post on Facebook dedicated to our recent study on OpenAI (this).
And the answer is precisely that Google does not “exit its” partly because it already has it (it’s called LaMDA and we told it here) and partly because it makes extensive use of artificial intelligence in many, many of its services. We are the ones who notice it the least o we don’t notice at all.
CES 2023
Increased or annihilated? What if artificial intelligence invades Hollywood
by our correspondent Emanuele Capone
Sundar Pichai’s speech in Davos
Be that as it may, in Mountain View they know very well that the question is circulating and that everyone expects some of their moves in this field: il CEO, Sundar Pichaihas already hinted that the company plans to unveil “a lot of things” in the field of AI applied to language in the course of 2023. And Pichai himself, speaking at the Davos Forumrecalled the “7 Ways Google Uses AI to Solve Big Social Challenges”.
Predicting floods
The first point touched concerns a an aspect we already wrote about on Italian Tech at the end of 2021: “Our flood forecasting program (called Flood Hub) is active in dozens of countries, displays forecast information in searches and on Maps, sends notifications to mobile devices for warn people in danger and features detailed flood maps so people can see exactly what to expect at their location.”
Fire boundary detection
Something similar is done with wildfires: “We use satellite imagery to build AI models that can detect wildfire boundaries in real time, showing their position in searches (on Google, ed) and on Maps. Last year we applied our models to more than 30 wildfires in the US and Canada, helping to inform residents and firefighters, and we recently extended the project to Mexico and parts of Australia as well.”
Prenatal health monitoring
Healthcare is another field where Google AIs they have been active for some time (here some examples from over 3 years ago): the most recent news is that “we are collaborating with Northwestern Medicine to develop and test artificial intelligence models that allow operators with minimal training and in contexts with limited resources to accurately identify potential problems and risks (related to childbirth and pregnancy, ed), such as the position of the fetus”.
Detection of genetic variations that cause disease
Still in healthcare: “Through a partnership with PacBio, researchers are using our DeepConsensus technology to identify genetic variants that cause disease,” which “it will help scientists discover serious genetic conditionssuch as an elevated risk of breast cancer or pulmonary arterial hypertension.”
Pest control in crops
Another delicate and important area: “We are collaborating with InstaDeep and FAO to better detect locust invasions in Africa, so that we can implement control measures” and also “we supported Wadhwani AI in India to create an app that helps identify and treat infestations of pests, with a 20% reduction in pesticide sprays and a 26% increase in profit margins for farmers”.
Understand population changes with building data
A relatively new project it’s called Open Buildings, is “based on open source data and uses AI to interpret satellite images, identifying the positions and geometry of buildings”. What’s its purpose? “This information is useful for population estimation, urban planning, humanitarian response during a crisis, environmental and climate science.” This tool is available in Africa as well as in Bangladesh, Indonesia, Laos, Nepal, PhilippinesSingapore, Sri Lanka, Thailand and Vietnam.
Education and AI
Is the Italian school ready for the impact of ChatGPT?
by Francesco Marino
Helping people connect and communicate
Finally, the idea of Project Relate, an Android app that “helps people who have difficulty making themselves understood to communicate more easily with others: can transcribe speech to text, use a synthetic voice to repeat what has been said or communicate directly with the speaker’s Google Assistant to complete a variety of tasks, from playing a song to turning on the light in the house. Remember a little the helpful voice navigation for the hearing impaired that is present in Mapsand is a decidedly concrete (and useful) application of the most advanced technology.