As Zuck Races to Build Godlike AI, Women and People of Color Aren’t Invited

Mark Zuckerberg has a new mission: set up an elite researcher, engineer and AI veterans, which can learn a form of Ai, Google, Antropic, Apple and more. This new unit, called Meta Supervise Labs (MSL), is given the task as the most powerful artificial intelligence in the world has seen so far.

Tech World calls him “a ghost group.” However, it is difficult not to notice that it is missing: Diversity.

Zuckerberg has been confirmed by 18 names and media news that has so far, only a woman. The list does not have black or Latin researchers. Most team members are men who participate in elite schools and are working in the best Silicon Valley firms. Many of the Asian origin

Here’s a partial list of new rents:

Alexandr Wang (CEO and Head AI Officer)
NAT FRIEDMAN (colic, former GitHub CEO)
In a hurry
Shuchao bi
Huiwen Chang
Ji lin
Joel Pobar
Jack Rae
Johan Shalkwyk
Pei sun
Jiahui Yu
Shengjia Zhao
Bad
Daniel is rude
Lucas Beyer
Alexander Kolesnikov
Xiaohua zhai
Ren Hongyu.

They are bright. This is under question. However, a similar piece is also truncated: the same institutions, the same networks, the same worldview. And this is a serious problem when building a strong thing as a super manager.

What is a superfish?

SuperTelence is the AI system that is smart people think, problem solving, problem-solving, creativity and even the smartest people in emotional intelligence. The best engineers can better write better code, the laws can make better analyzes better lawyers and manage companies more effectively than experienced CEOs.

In the theory, a super suite can revolutionize AI medicine, solve climate change or eliminate the traffic forever. However, it can also fix business markets, deepen control, expanding social inequality or automate harmful biases, especially if it reflects the perspective of those built, can automate harmful biases.

Therefore, who is important in the room. Because people who compile these systems are placed in the algorithms that can run the great parts of the society one day the values, assumptions and life experiences.

Whose intelligence is being built?

AI reflects designers. History has already shown what happened when the variety is not taken into account. Risks are hypothetical from facial recognition systems that fail in dark skin tones in dark skin tones, racist, sexist or predominated content.

The AI, built by homogeneous teams, tends to repeat the blind spots of their creators. This product is a defect. The goal is to build something smarter than humanity, the scale of these flaws.

This is like a god program. If you want to do this, a better mankind, not just a narrow sliver, be sure to understand mankind.

Zuckerberg said little about the composition of the AI team. In today’s political climate, the “diversity” is often rejected as a turn or a “hearth”, several leaders want to talk about it. But there is the value of silence. And in this case, the cost may be a reconnaissance system that does not see or serve most people.

An ongoing warning

Meta, said he built AI for everyone. But his personnel options offer otherwise. Black or Latino team members and about 20 hur are only a woman, sending a message to send a message to the company, or not, is compiling by several selected several selected.

Then there is a problem: Can we trust this technology? When making key decisions on cars, these machines are important to make sure that it understands the full range of human practices.

Now, if we do not make variety of diversity in the AI, we can eat the most unequalities of the future operating system.