On December 2, 2020, Timnit Gebru tweeted that she had been “immediately fired” from her role as co-lead of Google’s AI ethics team.

Exactly one year later, she announced her new venture, met with much fanfare in the tech community: the Distributed Artificial Intelligence Research Institute (DAIR). The global institute will focus on centering the marginalized communities that are most vulnerable to the tech’s harms.

Why it matters: Through both research and—eventually—consulting, Gebru hopes to push the tech industry toward a more ethical approach to AI deployment.

To recap, in December 2020, tensions between Gebru—then co-lead of Google’s AI ethics team—and Google leadership reportedly came to a head after a dispute over her research paper on the dangers of large language models. Gebru was ultimately fired from Google.

• Google has disputed Gebru’s version of events.

DAIR's vision

With DAIR, Gebru has said she wants to use her past experiences in Big Tech to create an environment where researchers can openly communicate their findings about AI’s harms, and where research subjects are acknowledged or paid.

Since the tech world skews white and male, and AI research skews Western, DAIR will recruit staff from global communities that are underrepresented and underserved in the tech world—people who can help build beneficial AI applications that may not be created otherwise.

• For example, DAIR is working on a public data set and paper showing how South African apartheid still affects land use in the country today.
• An AI analysis of aerial images suggested that between 2011 and 2017, in a region densely populated with poor people, most vacant land was turned into wealthy residences.

No strings attached

DAIR has already raised $3.7 million in funding from investors including the Ford Foundation, the MacArthur Foundation, and the Rockefeller Foundation. But moving forward, to help cement its independence, the institute will increasingly look to earn operating income from consulting work in AI ethics.

“The same big tech leaders who push out people like me are also the leaders who control big philanthropy and the government’s agenda for the future of AI research,” Gebru wrote in an op-ed for The Guardian. “If I speak up and antagonize a potential funder, it is not only my job on the line, but the jobs of others at the institute.”

Looking ahead: Gebru wrote that a system of checks and balances, as well as dedicated government funding for independent AI research, is needed to help distribute power in the tech world—e.g., “alternatives to the hugely concentrated power of a few large tech companies and the elite universities closely intertwined with them.”

Click here to read the full story.—HF

As the US chip-making industry clambers toward the close of a difficult year, it’s holding out hope for the House to pass a bill that would provide it with $52 billion in funding, Axios reports.

• The US has invested $0.00 in semiconductor manufacturing incentives over the last two decades, per the Semiconductor Industry Association.
• For comparison, China has spent around $50 billion on incentives over the same time frame, while South Korea has spent between $7 billion and $10 billion.

Semiconductor execs told Axios it now costs 30%–40% more to make chips in the US than in many of the Asian countries that have invested in the industry over the years.

For context: The United States Innovation and Competition Act (USICA) passed the Senate Back in June. In addition to the semiconductor funding, the bill is heavily focused on competing with China.

• Beijing has said the bill is based on “cold-war thinking” and that it will retaliate if it passes.

Last week, Commerce Secretary Gina Raimondo called on lawmakers to make USICA happen, and House Speaker Nancy Pelosi said she’d bring the relevant parties together to discuss the bill.

• Axios reports that the bill will likely become law eventually, but the window of time to pass it before the year’s end is slim.

Looking ahead…Even if the funding is approved tomorrow, it’ll take some time for the effects to be felt in the chip-making supply chain. Building up chip-making capacity is a labor-intensive endeavor, as we explained here, and as is evidenced by Intel’s two new plants. The company broke ground in September, but doesn’t expect the plants to be operational until 2024.

Click here to view on-site.—DM

November, like the 10 months of 2021 that preceded it, was another blistering month for venture funding. In total, companies raised $65 billion last month, per Crunchbase—the highest single-month total ever.

• For context: Global venture funding is already at $580 billion on the year—compare that to the total of $335 billion across all of 2020.

Let’s spotlight three of the emerging-tech companies that raised money last month.

Sierra Space: This company is laser-focused on low-earth orbit (LEO), and raised a $1.4 billion round last month, bringing its total valuation to $4.5 billion. You may remember Sierra Space as the company working on an interstellar bouncy house for Orbital Reef, Blue Origin’s planned private space station.

Ionity: The BMW-Daimler-Ford-Volkswagen-Hyundai fast-charging-joint-venture *deep breath* raised €700 million (~$790 million) last month. With the cash, it plans to grow its network of EV fast-chargers to 7,000 across Europe by 2025. Right now, it has 1,500 chargers in 24 countries.

Helion: Helion is a nuclear-fusion startup that raised half a billion dollars last month, with performance incentives that could net it another $1.7 billion. (Provided it delivers on the hitherto impossible scientific feat of generating net power from a fusion reactor, that is.)

• At the time, this was the largest-ever nuclear-fusion deal, but things move quickly these days—another fusion startup, CFS, raised $1.8 billion on Dec. 1.

While we’re here: Crunchbase just released an end-of-year report on diversity in VC funding.

Click here to view on-site.—DM

Stat: $26.9 billion worth of NFTs have been sold this year, but “a very small group of highly sophisticated investors rake in most of the profits from NFT collecting,” per Chainalysis.

Quote: “I think this next 12 to 24 months is a really big deal.”—Jonathan Levy, chief commercial officer at charging company EVgo, on the electrification of heavy-duty vehicle segments, to Emerging Tech Brew

Read: Tech analyst Benedict Evans published his annual tech trends deck.

Actually eventful events: Hopin’s all-in-one event management platform makes planning, producing, and re-living virtual events as easy as pie. Enable engaging interactive capabilities, make event content evergreen, and design on-brand virtual experiences that are just as exciting as in-person. Start here.*

^{*This is sponsored advertising content.}

We’ve seen two record-breaking nuclear-fusion funding rounds in the last month. How well do you know the elusive energy source?

Click here to take the quiz.

A new preprint from researchers at the University of Waterloo and Darwin AI, a computer vision startup, claims to have found a more efficient way to spot factory defects via algorithm. (H/t Import AI)

Enter: TinyDefectNet, a neural network trained to thrive in the fast-paced factory setting.

• The researchers say it can detect defects with the same accuracy as the state-of-the-art benchmark, but at “11x lower computational complexity.”
• In other words, it can reliably spot irregularities while causing less of a drag on factory production.

Catch up on the top Emerging Tech Brew stories from the past few editions:

• Batteries, the key component of EVs, keep getting cheaper. But that rate is slowing
• A primer on digital twin technology
• Fleets of electric buses, trucks, vans, and cars might soon help bolster the grid