A lot has happened this week that deserves attention in the AI room. The guard wrote an article with GPT-3 and again showed that no matter what OpenAI paid to train and build the language model, the free marketing could be worth more. After Amazon lost appeal against the JEDI cloud contract with the Pentagon, Amazon appointed Keith Alexander to its board – the man who oversaw the National Security Agency mass surveillance uncovered by Edward Snowden in 2013. And Portland passed the strictest bans on facial recognition in US history, banning government and corporate use of the technology.
However, AI Weekly tries to break into the zeitgeist and highlight one or the greatest events in people̵
In the Bay Area on Wednesday, the density of the smoke effectively blocked the sun, casting a deep, dark orange or red light. The eerie landscape drew comparisons Blade Runner 2049including sci-fi scenarios.
Someone used Bladerunner 2049 music for drone footage of San Francisco, and at first I didn’t know whether to be amazed or horrified. It is very terrible. pic.twitter.com/XQTv4qrE93
– Omar Jimenez (@OmarJimenez) September 10, 2020
On that day, the temperature was forecast to hit the high 80s or 90s, but temperatures dropped to around 60, representing a drop in heat of nearly 30 degrees in a single day. Street lights stayed on for most of the day and disoriented people felt like they were in the middle of the day at night. For some people who live here, the transition from San Francisco to Mars was a breaking point, a stressor after weeks of smoke and a heatwave alongside the ongoing battles against COVID-19 and making progress toward economic recovery and racial justice.
A quick look back: The August fire in Northern California is now the largest in the state’s history. CalFire said in its daily report today that 26 times as many acres were burned this year compared to the same period in 2019. At one point on Wednesday, the National Weather Service Bay Area tweeted that conditions were “beyond our models.” In Oregon, Governor Kate Brown said yesterday that 900,000 acres burned in three days. In a typical year, she said, the state will lose 500,000 acres to fire. Today, the unhealthy air quality spanned the entire west coast of the United States from San Diego to Seattle. On Friday, Portland had the worst air quality in the world. Salt Lake City had historically poor air quality in neighboring states, and Denver residents were asked to stay indoors.
The destruction and health risks for people caused by natural disasters are only one way of recognizing the consequences of climate change. The global environment is also under pressure: The World Wildlife Fund said in its annual Living Planet report, which monitored 21,000 species of animals, that the wildlife population has declined by nearly 70% over the past 50 years. This loss of biodiversity poses a threat to global food supplies.
So what can we do about it? There are major initiatives underway to apply machine learning to address the problem, such as the AI of climate change. The group of AI researchers is investigating solutions to climate change and related problems such as human dislocations or food insecurity. Over the past month, group members discussed the types of startups that can fight climate change. The group compiles a wish list for data sets from researchers to inform about what kind of data researchers want to train and solve problems.
In June 2019, researchers from more than a dozen organizations teamed up to publish a paper with more than 800 references that attempts to cover the multiple ways machine learning can help combat climate change. The full paper and interactive summary focus on practical models for electrical systems and smart cities, as well as long-term projects with uncertain implications like CO2 sequestration or the development of a planetary control system.
At an AI workshop on climate change held at the NeurIPS conference last year, researchers spoke about the possibility of making AI a carbon-free industry and the kind of cultural changes needed to enable the machine-based community Learning to pay more attention to climate change. Pocket calculators for determining the carbon footprint of a machine learning model were also distributed at the conference. Another AI workshop on climate change is to take place in December at NeurIPS.
There’s also the work of WattTime, a nonprofit that is reducing a household’s carbon footprint by automating when electric vehicles, thermostats, and appliances are active based on when renewable energy is available. Algorithms to determine these times are trained using data from the EPA’s continuous pollution monitoring system. The technology is currently available in California, where about 33% of its electricity is generated from renewable sources today as part of the self-generated incentive program, WattTime developer Gavin McCormick told VentureBeat.
“Nobody knows about the US Continuous Emissions Monitoring System, but it’s been around since the 1970s. That’s why companies like me can write ever more sophisticated AI algorithms to integrate more renewable energy and do what we do,” McCormick told VentureBeat in a phone call.
Last year, WattTime received a grant from Google.org’s AI Impact Challenge to see if Computer Vision could track emissions from power plants outside the United States using satellite imagery. In July, WattTime co-founded with nine organizations and Al Gore Climate Trace, a group that aims to track emissions in key economic sectors around the world such as power plants and shipping. Climate Trace aims to make this data available to the public by June 2021, until the next round of international climate negotiations.
The idea that technology could change the world grew up, an idyllic dream. There has been a fair share of disillusionment over the past few years, partly due to startups solving non-existent problems, mass surveillance, general lack of funding for various startup founders, and too long a list of violations from big tech companies to list here.
But if you are the one person who feels helpless in the face of recent disaster and wants to do something to change things, numerous projects (including the one above) require volunteers. According to a poll published by Yale Climate Change Communication in April, people are ready to join in efforts to enforce action by elected officials to combat climate change.
The word apocalypse came up a little too often for my taste this week. It’s easy to sense that things are awful, messy, and out of control – because they are – but no one should believe that there is nothing they can do to change anything. Technology alone will not save us. People need to vote for elected officials whose policies take climate change seriously, and individuals can take action.
Do you know of other projects at the interface of climate change and machine learning? Send news tips to Khari Johnson and Kyle Wiggers, as well as AI editor Seth Colaner – and be sure to subscribe to the weekly AI newsletter and bookmark our AI channel.
Thank you for reading,
Senior AI Staff Writer