Good morning. In what I am sure is a ground-breaking mixup, our AI for Good story is coming up last this morning.

I spoke with IBM about the latest project — in partnership with the UNDP — to come out of its Sustainability Accelerator.

— Ian Krietzberg, Editor-in-Chief, The Deep View

In today’s newsletter:

OpenAI launches independent safety committee

OpenAI on Monday said that it is converting its Safety and Security Committee — which was established in May amid mounting safety concerns — into an independent board oversight committee. 

This conversion is one of five recommendations that the Safety and Security Committee has made, all of which OpenAI is adopting. 

The details: The newly independent committee will be chaired by Zico Kolter, director of Carnegie Mellon’s machine learning department. 

The company said that the committee will also “exercise oversight over model launches, including having the authority to delay a release until safety concerns are addressed.”

In terms of the other measures OpenAI is taking, the company promised to enhance its cybersecurity measures, increase transparency and collaborate more with external organizations. 

The context: In May, OpenAI began bleeding safety researchers — at the time, Jan Leike, who formerly co-led the company’s Superalignment team, said that safety was not a priority at OpenAI. The company has since gained a few vocal critics in the form of former employees turned whistleblowers. 

Ready to become a leader in the AI revolution?

ZapConnect, the most anticipated automation event of the year, is your chance to dive deep into the future of AI and automation. Packed with insights and hands-on sessions, this event will give you the knowledge to elevate your workflows and lead in AI innovation.

Why ZapConnect is a Must-Attend:

And the best part? It’s all happening for free on September 26th.

Register today for free.

Intel’s (possible) turnaround

Intel, once the clear king of the semiconductor kingdom, has been struggling as of late.

Nvidia took all the oxygen out of the room when the AI craze began; shares of Intel have tumbled around 60% throughout this year as the company has struggled to keep up. 

What happened: The company this week unveiled a bunch of new arrangements that have given beleaguered investors something to cheer about. 

Intel also announced a shift in its structure that would make its foundry businesses an independent subsidiary that exists within Intel. The intention behind the separation is to assure potential customers — some of whom might be in competition with Intel — that the company is a neutral entity. 

The stock began to climb on Monday and Tuesday, though at around $21.50 per share, the stock is still far off its July levels of around $30, which it shed in the wake of a bleak earnings report last month. 

CEO Pat Gelsinger said in a statement that “this is the most significant transformation of Intel in over four decades.”

If you want to get in front of an audience of 200,000+ developers, business leaders and tech enthusiasts, get in touch with us here.

Microsoft, G42 setting up new AI centers in Abu Dhabi 

Microsoft is launching a new center for responsible AI — co-founded and co-founded by the Abu Dhabi-based AI developer G42 — in Abu Dhabi. 

The details: The center, launched with the endorsement of Abu Dhabi’s Artificial Intelligence and Advanced Technology Council (AIATC), will advance responsible AI standards in the Middle East and the Global South. 

At the same time, Microsoft announced that it is launching a new AI for Good lab in Abu Dhabi, its first in the Middle East. The center will work with organizations and nonprofits to address “key economic and societal challenges” across the Middle East and Africa. 

The firms said that they have been executing their partnership with the “support and oversight” of both the U.S. and UAE governments. 

AI for Good: IBM and UNDP launch energy equity program

For the past two years, IBM and the United Nations Development Programme (UNDP) have been working on an AI-based project intended to ensure energy equity in developing nations. 

The project, developed through IBM’s Sustainability Accelerator, involves the deployment of two new models, which IBM and the UNDP unveiled on Tuesday. 

The details: The first is an electricity access forecasting model, which is publicly accessible here. 

The second model is a clean energy equity index, which is publicly accessible here. The model — which IBM calls a “first-of-its-kind” system — combines geospatial analytics with social, environmental and economic factors, such as greenhouse gas emissions, education and relative wealth levels. 

IBM launched its sustainability accelerator — a pro-bono program that brings IBM’s tech to nonprofits and government agencies with the goal of advancing global sustainability efforts — in 2022. 

Why it matters: The aspiration of this program, according to Michael Jacobs, the creator of IBM’s sustainability accelerator and the lead of its program with the UNDP, is “to better equip UNDP country offices that actually are the interface between, oftentimes, the whole of the UN and national governments, and then policymakers or other practitioners to get access to insights around the energy picture for their country.”

Jacobs told me that the clean energy equity index is vital for enabling organizations to prioritize the most urgent and impactful opportunities for greener energy mixes, helping ensure that clean energy not only comes online, but is actually accessible to people. 

An element to this is that it is quite energy-intensive to both construct and operate AI models. Sometimes, this energy requirement is so intensive as to negate any sustainability progress a given model is able to achieve. A recent Guardian analysis found that, even as Big Tech has admitted to soaring, AI-driven spikes in its energy usage and carbon emissions due to AI, actual data center emissions are likely 662% higher than Big Tech claims. 

In response to this dichotomy of AI’s energy appetite, Jacobs said that the approach has to be all about efficiency. 

“Here, the eye has always been around efficiency, both in terms of workloads for compute, building something that's affordable, something that's efficient to run for the UN itself,” he said.

I’ve found the idea of this recurring “AI for Good” section to be super important, not just to demonstrate how the tech is helping, but to also demonstrate how the tech is not helping. 

That kind of intentionality of purpose feels markedly different from much of the consumer-facing generative AI products out there that make big promises, generally don’t live up to them and largely move far too fast, without considering the social and ethical impact of that speed. 

This is the fundamental difference between AI for good, and everything else. 

“The accelerator is a social impact program at its core,” Jacobs said. “So we care about environmental sustainability, but ultimately, we are prioritizing that work to benefit people on the ground.”

🤔 Your thought process:

Selected Image 2 (Left):

Selected Image 1 (Right):

💭 A poll before you go

Thanks for reading today’s edition of The Deep View!

We’ll see you in the next one.

Here’s your view on Microsoft Copilot:

A third of you said you will never use Copilot. 22% said it’s as good as advertised and another 22% said you hate it. 12% said it’s okay, but isn’t really a time-saver.

As good as advertised: