人工智能將平等帶進(jìn)職場

埃米莉·采蓋爾斯基

Inequality in the workplace has long been a hot topic in the corporate world， but gender economist Katica Roy is actively turning that conversation into action. After fighting to be paid equitably twice—and winning—Roy was on a radio show when asked if she thought the pay gap would be closed in her lifetime.

“I said， ‘Well not until we make it an economic issue. And then I thought， oh， I think I can solve that.” Roy explained to Worth CEO Juliet Scott Croxford， during the first day of the weeklong Techonomy Virtual： Reset + Restore conference. “So， that was my entry into entrepreneurship and really looking at gender equity as a massive economic opportunity rather than only a social issue.”

This realization led Roy to found Denver-based Pipeline Equity， an award-winning SaaS company that leverages artificial intelligence （AI） to identify and drive economic gains through gender equity.

“The idea behind [Pipeline and its application of AI] was： If we could change decisions that are made in organizations， the human capital decisions， we could actually make gender equity a reality in our time in our lifetime， rather than the hundred or so years that the World Economic Forum （WEF） talks about.” Roy said. “But also， we could essentially augment the decision making. We could also change the narrative from gender equity as this social issue to an economic opportunity. The thing that we also saw more broadly in the marketplace was that there was an increasing number of CEOs who were committed to gender equity and have made these public commitments. Yet， there was really a difference between making a commitment and living a commitment， really the difference between making that commitment and actually the employee experience of that commitment.”

“So， if we could change the way that the decisions are made，” she continued， “then we could actually catapult that opportunity forward.”

Of course， the WEF pay gap closure predictions were made before the COVID-19 pandemic hit， which Roy called a “she-session，” because women have lost the majority of jobs during this particular economic downturn.

“With remote work and with people being at home with children， most caregiving and unpaid work falls to women，” Roy explained. “Weve seen that theres about 40 percent of U.S. employees [who] dont have sick leave. So， they have to choose between their health and going to work. And most of that actually falls on women… this particular economic downturn has really highlighted all the cracks that were already sitting underneath the system.”

But while this crisis has been a “giant step backward” for equality， “at the same time， what we are seeing is this opportunity to actually use artificial intelligence and use people analytics solutions to catapult us forward，” Roy said.

With the move to remote work， there has been a lot of talk about the flexibility of such situations being good for gender equity， but Roy warned that while working from home could help keep more women in the workforce， it could have a negative impact on female promotions.

“At this moment， theres the opportunity to actually leverage artificial intelligence to ensure that we are not only continuing to make progress on gender equity， but actually catapulting us forward，” Roy explained. “And for companies， that is of particular importance because in the 2008 Great Recession， what we saw was companies that put equity at the core of their crisis management strategy actually increased the velocity of their recovery. So， for companies that are looking to recover， and recover more quickly， from this particular downturn， gender equity is important.”

But to close the gender pay gap， we cant start by talking about pay. “Pay is the symptom， its not the disease，” Roy said. “In other words， pay is the quantitative value that you place on your talent. But the actual value happens before that in performance and potential.” And thats where Pipelines platform steps in.

“We are augmented decision-making，” Roy， who also acts as Pipelines CEO， said. “So much like you would use Google Maps or ways to get from point A to point B， we do the very same thing for companies human capital management decisions. Essentially， its their data， our algorithms. We attach to their HR systems when theyre going to make a decision across these five pillars of talent， which essentially are kind of the five big buckets of talent decisions you make： hiring， pay， performance， potential and promotion. We actually intercept those decisions， essentially audit them. And then if theres any inequity， make a recommendation.”

Using natural language processing， Pipelines platform reads through performance reviews and calls out any biased phrases. “Weve found that on average， women are underrated 4 percent of the time， and that actually impacts their ability to be in the succession pipeline， as well as their pay. So thats what we make possible. Just to give you one more quick stat： The average Fortune 500 company has 60，000 employees. And what we have discovered is that there are really three key decisions that they make across those employees each year， which are performance， potential and pay. So， thats 180，000 opportunities for the average Fortune 500 company to move toward equity each and every year. Thats what we make possible.”

According to Roy， in the next five years， we could reduce the time it will take to tackle gender equity， which is currently estimated at 257 years. “We could actually shorten that quite a bit，” Roy said. “At this moment， we have the opportunity to embrace AI as a tool to achieve gender equity.”

職場中的不平等一直是企業(yè)界的熱門話題，而性別經(jīng)濟(jì)學(xué)家卡蒂察·羅伊正在積極地將這一討論轉(zhuǎn)變?yōu)樾袆?dòng)。兩次為同工同酬而斗爭并且都大獲全勝之后，她在一次電臺(tái)節(jié)目做客時(shí)被問到，男女薪資差距在她有生之年是否會(huì)消除。

“我當(dāng)時(shí)的回答是，‘不把它變成一個(gè)經(jīng)濟(jì)問題就不太可能。然后我想，哦，我覺得我可以解決這個(gè)問題?！痹跒槠谝恢艿摹凹夹g(shù)經(jīng)濟(jì)虛擬大會(huì)：重啟+恢復(fù)”首天，羅伊這樣對(duì)價(jià)值媒體公司首席執(zhí)行官朱麗葉·斯科特·克羅克斯福德解釋，“就這樣，我走上了創(chuàng)業(yè)之路，真正開始將男女平等視為一個(gè)巨大的商機(jī)，而不僅僅只是一個(gè)社會(huì)問題?！?/p>

認(rèn)識(shí)到這一點(diǎn)后，羅伊便創(chuàng)建了總部位于丹佛的Pipeline Equity。這是一家獲獎(jiǎng)的SaaS公司，充分利用人工智能，通過性別平等舉措，找到經(jīng)濟(jì)增長點(diǎn)并促進(jìn)實(shí)現(xiàn)經(jīng)濟(jì)收益。

“我（創(chuàng)建Pipeline并運(yùn)用人工智能）的想法是，如果我們能夠改變公司做出的決策——也就是人力資本決策，那么我們?cè)谟猩昃湍茏屇信降瘸蔀楝F(xiàn)實(shí)，就不用等到世界經(jīng)濟(jì)論壇所說的百來年后?！绷_伊說，“而且，我們可以從根本上加強(qiáng)決策過程。我們還可以改變提法，把男女平等從社會(huì)問題變?yōu)樯虣C(jī)。我們?cè)谑袌錾细鼜V大范圍內(nèi)看到，越來越多的首席執(zhí)行官致力于實(shí)現(xiàn)男女平等，并已做出公開的承諾。但是，做出承諾和遵守承諾是兩回事，做出的承諾和員工的感受到那份承諾又是兩回事?！?/p>

她繼續(xù)說：“所以，如果我們能夠改變做決策的方式，就能抓住機(jī)會(huì)、突飛猛進(jìn)。”

當(dāng)然，世界經(jīng)濟(jì)論壇關(guān)于消除男女薪資差距的預(yù)測是在新冠疫情暴發(fā)前作出的，羅伊將新冠疫情造成的沖擊稱為“女性期”，因?yàn)樵诒敬谓?jīng)濟(jì)衰退期，大多數(shù)失業(yè)者是女性。

“由于遠(yuǎn)程工作，加上孩子也都待在家中，照看孩子和無償工作主要就落在婦女身上了?！绷_伊解釋說，“我們已經(jīng)看到，大約40%的美國員工沒有病假。因此他們不得不在健康和上班之間做出選擇。而面臨這一選擇的實(shí)際上主要是婦女……這次經(jīng)濟(jì)衰退，讓我們社會(huì)體系中早已存在的裂痕真正凸顯出來?！?/p>

羅伊表示，雖然疫情危機(jī)對(duì)于男女平等來說是一次“大倒退”，但“與此同時(shí)，我們也看到了使用人工智能和人員分析解決方案來幫助我們大步邁進(jìn)的機(jī)會(huì)”。

由于工作轉(zhuǎn)為遠(yuǎn)程居家辦公，很多人討論說，這種情形的靈活性有利于男女平等，但羅伊警告說，雖然居家工作可以使更多女性留在職場，但它對(duì)女性職場晉升卻有負(fù)面影響。

羅伊解釋道：“目前來看，我們有機(jī)會(huì)運(yùn)用人工智能，以確保我們不但在男女平等上繼續(xù)取得進(jìn)步，而且是大踏步向前。而對(duì)于各大公司來說，這一點(diǎn)特別重要，因?yàn)樵?008年經(jīng)濟(jì)大蕭條期間，我們看到的是，將性別平等作為危機(jī)管理策略核心的公司復(fù)蘇速度更快。因此，在這一次的經(jīng)濟(jì)衰退期，對(duì)于想要復(fù)蘇、想要更快地復(fù)蘇的公司而言，男女平等很重要。”

但是，要想消除男女薪資差距，我們不能從談?wù)撔劫Y開始?！靶劫Y只是癥狀，并不是疾病本身。”羅伊這樣說，“換言之，薪資是對(duì)才能的價(jià)值量化。但一個(gè)人的表現(xiàn)與潛力所產(chǎn)出的實(shí)際價(jià)值是先于這一量化價(jià)值的?！倍@正是Pipeline的平臺(tái)切入點(diǎn)。

兼任Pipeline 首席執(zhí)行官的羅伊說：“我們是增強(qiáng)版的決策過程。就像你使用谷歌地圖或者其他方式從一個(gè)地點(diǎn)出發(fā)到另一個(gè)地點(diǎn)一樣，我們?yōu)楣救肆Y本管理決策所做的是完全一樣的事。基本上是公司出數(shù)據(jù)，我們出算法。公司招聘人才時(shí)需要在招聘、薪資、績效、潛力和晉升等五個(gè)重要方面做出決策，這時(shí)，我們接入他們的人力資源系統(tǒng)，攔截這些決策，基本上是對(duì)這些決策做必要的審核。如有不平等情況出現(xiàn)，則會(huì)提出建議?！?/p>

Pipeline平臺(tái)使用自然語言處理技術(shù)，檢讀績效評(píng)估報(bào)告，挑出帶有偏見的措辭?！拔覀儼l(fā)現(xiàn)，平均而言，對(duì)女性的評(píng)估低了4%，而這實(shí)際上影響了她們后續(xù)的升職加薪。是我們發(fā)現(xiàn)了這些有待改進(jìn)的方面。再快速瀏覽一條數(shù)據(jù)信息：《財(cái)富》雜志評(píng)選的世界500強(qiáng)公司平均每家有6萬名員工。而我們發(fā)現(xiàn)，他們每年對(duì)員工做出的其實(shí)有三項(xiàng)核心決策，即績效、潛力以及薪資。也就是說，對(duì)于每一家世界500強(qiáng)公司來說，每年有18萬個(gè)推進(jìn)男女平等的機(jī)會(huì)。這些機(jī)會(huì)是我們找到的?！?/p>

在羅伊看來，在接下來的五年間，我們能逐步減少實(shí)現(xiàn)男女平等所需的時(shí)間——目前估計(jì)需257年。“我們能將這一時(shí)間縮短許多?！绷_伊說，“而現(xiàn)在，我們有機(jī)會(huì)應(yīng)用人工智能這一工具實(shí)現(xiàn)男女平等?！薄?/p>