OpenAI Spends $1 Million to Hire More Black People to Promote Racial Equality

Recently, the "Black Lives Matter" movement has spread to the technology circle, and many technology companies have also made statements on the matter. OpenAI recently announced a decision to invest $1 million to vigorously recruit black members to help achieve racial equality. 

Recently, the racial equality movement in the United States has gained momentum, and many technology companies have joined in. After Google and Go announced the abolition of blacklist/whitelist naming, OpenAI also came up with the latest policy to support equality.

OpenAI spends $1 million to promote racial equality

On June 9, OpenAI posted a long article on Twitter, saying that it hopes to do something to helpBlack equity.The final decision was to expand the Scholars Program.Recruit a large number of members of the black community and invest $1 million to provide educational resources and guidance to minority groups.

The full Chinese translation of his tweet is as follows:

At OpenAI, we’ve been discussing what we can do to help black people achieve equality. Now, we’re announcing thatHelp will be provided through direct donations and expansion of our Scholars Program.

We have committed $1 million to this program to provide educational resources and mentorship to underrepresented minority groups in the AI industry.

We are currently in the process of significantly increasing the number of scholars, and we are heavily recruiting members of the black community. We will be excited if all the spots are filled. We will share details about the next class of scholars in the coming weeks.

At OpenAI, our goal is to ensure that in the future, advanced AI systems are built to benefit everyone. This means we also need to work harder to diversify our workforce. So far, we haven’t been doing a good enough job of this.

It’s not enough for a company to be non-racist or to tweet “Black Lives Matter.” We want to be actively anti-racist, and to do that we need to make active corrections in the way we share spaces and opportunities. We’re committed to doing this work.

Under this tweet, some netizens said,This is just a PR move with no real meaning:

Some netizens were also happy about this move.Think OpenAI has taken a welcome step:

This year marks the fifth anniversary of the founding of OpenAI. In 2017, the AI developed by OpenAI defeated the world's top player Dendi in the 1v1 competition of the Dota2 International Invitational, and became famous.

The company later released a number of research results, which attracted widespread attention in the industry. In July 2019, OpenAI received a $1 billion investment from Microsoft to jointly develop cloud AI technology.

This Silicon Valley star company has grown stronger and stronger, and its team has grown from the initial 9 people to more than 100 people.

However, we noticed that the four senior executives on the team were all white.Judging from the latest public team photo,The vast majority are white or Asian.A search on LinkedIn can also verify this.

So, does this lack of diversity in the team make technology racist?

We all know that as AI is increasingly used, more issues regarding bias are exposed.

Two years ago, a study by MIT and Microsoft demonstrated that facial recognition software can identify lighter-skinned men with over 99% accuracy;But for women with darker skin, the software could be wrong up to a third of the time.

An engineer from LinkedIn also mentioned,Previously, Amazon's artificial intelligence recruitment system gave priority to men when recommending candidates;

When LinkedIn's intelligent algorithm recommends jobs to users, it mostly recommends high-income occupations to male job seekers, but rarely recommends high-income occupations to women.

Therefore, leaving aside the issue of training data for the moment, just speaking of developers, if all the developers behind a technology are men, or racists,It is inevitable that some subjective colors or stereotypes will be introduced.

In fact, OpenAI has already noticed this problem.Every year, the company discusses issues such as bias, safety, and dangers of AI technology.

Some time ago, OpenAI mentioned in a blog post about improving AI verifiability,Developers should set bias and safety rewards for AI systems to increase incentives and scrutiny of AI systems.

This time, OpenAI directly and publicly announced that it would recruit more black members, and it seems that its attitude towards eliminating racial and technological bias is still very determined.

But as to whether they really do what they say or it is just a public relations move to cater to the public, only time will tell.

OpenAI has money to burn and is not short of money 

By the way, what does $1 million mean to OpenAI? Since its inception, OpenAI has been developing projects that require huge costs.

Take just one of the studies, GPT-2, released in March 2019, a model with 1.5 billion parameters, trained using 256 TPU v3s.That costs $2,048 per hour.

According to a previous report, in 2016, when OpenAI had only more than 50 employees, its salary and benefit expenses reached 7 million US dollars.

Therefore, donating $1 million is not a luxury for OpenAI.

-- over--