6 unconventional ways AI is being used to help the planet

Welcome to The Dispatch! We are the newsletter that keeps you informed about AI. Each weekday, we scour the web to aggregate the many stories related to artificial intelligence; we pass along the news, useful resources, tools or services, technical analysis and exciting developments in open source. Even if you aren’t an engineer, we’ll keep you in touch with what’s going on under the hood in AI.

The unconventional ways that AI is helping the planet - from cheese to steel

From Fast Company: The transformative potential of generative AI is not limited to generating text or images - it’s making a tangible impact in the battle to combat climate change. An article from Fast Company discusses six pioneering ways in which generative AI is currently aiding in climate mitigation, ranging from innovating plant-based food alternatives to optimizing the carbon footprint in steel production and even facilitating greener urban planning.

More details:

• Startups like NotCo and Climax Foods are using generative AI to create more realistic plant-based meat and dairy products by rapidly testing ingredient combinations. This could help reduce livestock emissions.

• Generative AI is being used to design novel EV battery chemistries that avoid scarce materials like cobalt. The tech can suggest new material combinations in months rather than years.

• Researchers are applying generative AI to speed up the discovery of catalysts needed to produce clean fuels from renewable electricity.

• In architecture and construction, AI could analyze data from past projects to optimize building efficiency and materials. Buildings are responsible for as much as 40% of climate emissions.

Takeaways: The range of emerging use cases for AI continues to grow and shows the creativity underway in leveraging AI for environmental benefits. The technology's use in optimizing steel recycling mirrors shifts in waste management practices from the early 2000’s. Back then, advancements in sorting and recycling technologies led to a massive reduction in landfill waste. Here, AI has the potential to play a similar role but in a far more complex and urgent arena.

As we have repeatedly noted in this publication, generative AI comes with its own environmental costs. For now, it’s a delicate balancing act - let’s hope MIT’s light-powered dreams can put AI’s energy consumption concerns to rest sooner than later.

LinkedIn: 21x increase in job posts mentioning AI in less than one year

From LinkedIn: A new quarterly report from LinkedIn examines how AI is transforming the workforce. The report leverages the platform’s unique labor market insights (taken in aggregate from more than 950 million professionals on LinkedIn) to look at trends in AI job postings, skills, and sentiment among executives and employees. Surprise, surprise - the report finds AI adoption accelerating across industries, with key themes being the rise of generative AI and an increased importance being placed upon "people skills".

More Details:

• There has been a 21x increase in job postings mentioning AI technologies like ChatGPT since November 2022

• LinkedIn members adding AI skills to their profiles have grown 9x from January 2016 to June 2023 globally.

• 47% of US executives believe generative AI will increase productivity, and 92% agree people skills are more important than ever.

• Skills like flexibility, professional ethics and social perceptiveness are the fastest growing in-demand skills in US job postings since November 2022.

• Of all industries analyzed in the US, financial services stands out for having both the largest share of members with AI skills (at 0.9%, for perspective) and the fastest pace of adding those skills, with a 30x increased pace since January 2016.

Takeaways: Overall, this report paints an optimistic picture of AI's impact on work - increasing productivity and value-added activities rather than wholesale job loss. However, it also highlights the need for companies and workers to skill up on AI technologies and adapt roles. Just as past technological shifts led to new opportunities, organizations and professionals that embrace upskilling and evolve their capabilities will be best positioned to meet the future. For developers and technologists especially, actively expanding skills in generative AI will open new career paths .

The only claim OpenAI didn't seek to dismiss alleges direct copyright infringement, which it hopes to defeat later. From our perspective? The unauthorized use (or scraping) of copyrighted works for commercial gain is unacceptable - we hope that precedent gets set very clearly in these cases.

• IBM and Salesforce launch new AI collaboration

• Call of Duty will use AI to moderate voice chats

• U.S. bans sales of Nvidia's H100, A100 GPUs to Middle East over concerns of re-exportation to China

• (SynthID) Google tests watermark to identify AI images

• World Economic Forum: Why we need to be realistic about generative AI’s economic impact

The SDK handles infrastructure so developers can focus on building with apps. For example, a startup called Memorang used it to build an educational platform with AI-generated content, assessments, and study tools. The founder and CEO, Dr. Yermie Cohen, explained that Memorang was “built on the modern and evolving AI stack, including Vercel, much of which didn’t exist months ago.”

To address communication delays, How's team added "perception aware" capabilities so vehicles can gather real-time sensor data on others and adapt trajectories accordingly. This resulted in completely collision-free drone flights during testing. How is also creating models to predict pedestrian behavior using autonomous car sensors, reducing uncertainty. “The real goal is to improve knowledge. You're never going to get perfect predictions. You're just trying to understand the uncertainty and reduce it as much as you can.”

• (Google research) Modeling and improving text stability in live captions

• Watch Microsoft Copilot ‘Power Automate Process Mining’ in action

• Intel Habana Gaudi beats Nvidia's H100 in visual-language AI models: Hugging Face