Global Climate Crisis: Is AI a bane or a panacea?

The conversation about climate change has been ongoing since I was in high school back in late 90s. The term was “global warming” back then. While we have progressed to calling it with better names that accurately represent the phenomenon, it sometimes feels we’ve not moved much further in terms of action. Every year the scale and intensity climate change’s effects have been intensifying and we have been struggling to catch up. Advancements in tech have provided useful tools to combat some of the effects, but they have also accelerated emissions and resource consumption as a side effect. And when AI moved out of the research labs onto the commercial arena, it has supercharged the rate of consumption. For all its promise, are AI advancements going to veer us away from the cliff we are marching toward? Or is it going to push us off even faster?

Grim Reminders being met with global inaction

With most North America grappling with heatwaves, we’re witnessing yet another grim reminder of the effects of climate change. In the recent past, nature has been serving several such reminders - the once-in-200-year heatwaves in Africa’s Sahel region in April took thousands of lives in just four days; flooding in the arid middle east claimed several lives including sweeping away a bus full of school children; and the Canadian wildfires displaced over four thousand residents in British Columbia. And this is just a smattering of examples from the last few months. Climate driven tragedies have been piling up so fast that we have sadly become desensitized to its impact.

The conversation about climate change has been ongoing for decades now, while progress has been mostly incremental. Over 80% of the world’s energy still comes from fossil fuels. A trend that is predicted to last at least until 2050, driven from demand in developing nations. Economic and technological disparity is a commonly used argument for a lack of coordinated global effort to reduce fossil fuel dependency. But this argument doesn’t hold its water, as 10 of the 17 nations that’s executed little to no climate action happen to be developed economies. And, almost all of them are due to lack of policies and societal follow-through to align with the Paris agreement they had signed.

Tech behemoths struggle to put the genie back in the bottle

Tech has been a major contributor as well. The maturity of Cloud Computing has fueled power consumption and carbon emissions to even greater amounts in the last decade. And not to mention, the recent proliferation of AI has supercharged this phenomenon. Data centers that power the Cloud contribute to as much CO₂ emissions as the airline industry, and training a single AI model can emit CO₂ equivalent to five times the lifetime emissions of an average car.

While Cloud behemoths like Amazon have claimed to have achieved lofty goals like running 100% of data centers on renewable energy, the real number is not more than 22% according to their own employees. Microsoft has plans to become carbon neutral by 2030, but their emissions have been increasing over 25% year-over-year since 2020. Google has stopped calling itself carbon neutral as their energy consumption has doubled since 2019.

The nascent promise of AI

For all the hype, has AI really delivered anything tangible to combat the climate change it is rapidly accelerating? The answer is a mixed bag of small improvements. Most are geared to better react to disasters while a handful of ideas are emerging with real potential for long term impact.

For instance, NASA partnered with engineering firm Development Seed and used Machine Learning to cut down estimates of hurricane Harvey from 6 hours to mere seconds. That’s a huge leg-up for for disaster relief teams to act proactively. Google recently published a paper on using AI to predict floods as early as 7 days in advance, and providing this data to over 80 countries. A significant tool to reduce casualties in densely populated nations like India.

For a field that has only really caught on over the last year and a half, it is promising to see AI make an impact already. And there is no doubt its effect to combat climate change will only improve. But will the impact be substantial enough to really surpass the damage it’s causing the environment?

AGI versus human tendencies

Artificial General Intelligence (AGI) is the future of AI when systems one day mimic human-like capabilities of reasoning, creativity and intelligence. The optimistic view is AGI will offer breakthroughs across a wide range of fields that combat the roots of climate change with lasting impact. Such as accelerating the research of artificial photosynthesis, molecular nanotechnology and nuclear fusion. However AGI is still a few years away and needs several advancements in tech including commercial quantum computing which is still largely a research area.

Even if AGI materializes, humanity needs to address the friction posed by special interest groups. Organizations like American Petroleum Institute, National Association of Manufacturers and several coal companies have Supreme Court granted abilities to form Super PACs to lobby against clean energy initiatives. Firms like Electric Utilities have found ways to weasel in and introduce loopholes into final language of laws and weakening rheir enforcement.

With all the fog of war, major questions still loom large - will AGI’s impact be too little too late? Or will it be significant enough to offset the hurdles we are posing for ourselves? How many one-way doors will humanity have to cross as a result of climate change induced disasters before the sense of urgency reaches critical mass? Ultimately it doesn’t matter if AI is the panacea we are looking for, if the human gridlock cannot be broken.

AI Bites

Follow up from last week’s Deep Fakes

The proverbial cat strikes back in the cat and mouse game of the U.S Election sabotage, as the US Justice Department disrupted a Russian influence operation that used AI-enhanced fake social media accounts to spread pro-Kremlin messages in the US and abroad. The operation was conducted by a Russian private intelligence organization, backed by the Kremlin, aimed at influencing the upcoming US presidential election. The Justice Department collaborated with Dutch officials and seized multiple domain names and searched around 1,000 social media accounts to identify the operation was running from a server in the Netherlands. This effort marks the first US public accusation of a foreign government employing generative AI in influence campaigns.

US and EU Regulatory scrutiny in AI monopolization yields results

Amid increased regulatory scrutiny in the EU and US over Big Tech's AI investments, Microsoft has withdrawn from its non-voting observer role on OpenAI's board, and Apple has opted not to join. OpenAI will now update partners and investors through regular meetings. Microsoft's deputy general counsel cited board progress for the change. Apple, initially considering a role for Phil Schiller, also opted out. OpenAI plans to engage partners like Microsoft, Apple, Thrive Capital, and Khosla Ventures through meetings under CFO Sarah Friar. Despite stepping back from board roles, Microsoft remains a key partner with over $10 billion invested. Regulatory concerns likely influenced these changes, with investigations into Big Tech's influence in AI.

Claude Projects make agentic AI more accessible

Claude Projects AI is shaking up the automation game, offering tools to make your workflows smoother and more efficient. It promises capabilities that can streamline tasks by documenting workflows, setting up prompt sequences, and developing precise instructions. It’s especially useful for automating tasks like LinkedIn campaigns, significantly reducing human error and saving time. The AI also supports advanced features such as custom GPT models and synthetic test data generation, along with addressing common challenges like data accuracy and security; making it a powerful tool for enhancing productivity.