Silicon Valley’s other diversity problem
This week’s hot tech topic has been the Google Memo, an open document written by a Googler, James Damore, which has created a stir in Silicon Valley and beyond.
Much of the memo speculated on whether innate gender differences might be responsible for the lack of women in coding. Several of the author's comments were perceived as perpetuating gender stereotypes, leading to his being fired by the tech giant, already twitchy over claims of pay discrimination.
The issue of gender in digital tech warrants a dedicated discussion, especially since much of the popular press commentary tends towards the simplistic and emotionally-charged. (Suffice to say, there is evidence both for supply-side and demand-side problems - and lashings of confirmation bias all round).
For the moment, though, I want to concentrate on the other main claim of Damore’s memo: that Google has developed an authoritarian culture which will not tolerate dissenting voices, especially those from a conservative perspective. In Damore’s words: "Google’s left bias has created a politically correct monoculture that maintains its hold by shaming its dissenters."
Damore is not the first to have voiced these fears. A few months ago, the well-known venture capitalist Marc Andreessen spoke openly about what he saw as institutional political bias of Californian tech industry. "Silicon Valley is extremely left-wing", said Andreessen, before claiming that "99.99 per cent of Silicon Valley last year voted for Hillary Clinton."
Now, as a whole, California is clearly strongly pro-Democrat: in the last US election, around 62 per cent of the State voted for Clinton. In San Francisco, home of 1960s counterculture, this rose to 85 per cent.
However, there are increasing claims that political leanings within tech firms are even more pronounced, and that these political views are colouring their activities
Twitter, for instance, is regularly charged by conservatives of censoring right-wing users via banning, ‘shadow-banning’, manipulating hashtags, unequal enforcement of terms of service and preferential verification.
(In fairness to Twitter, evidence for this is largely anecdotal, and may in part be an incidental effect of its curation algorithms: since Twitter users are on average more left-wing than the general public, this may influence the type of content which is given prominence.)
Facebook has also come under fire for left-wing bias, particularly after former employees last year admitted ‘routinely’ suppressing conservative news and manipulating feeds on the platform, prompting a letter of enquiry from the US Senate.
Google, even before this week, had something of a reputation for political tilt, too. Following the 2012 US elections, statistician Nate Silver analysed itemised political donations from employees of tech companies. On average, these were about the same as for San Francisco county as a whole (that is to say, overwhelmingly Democrat). However, among Google employees, some 97 per cent of donations to main candidates favoured Obama rather than Romney. (Netflix was the only tech giant with a greater proportion).
While funds donated is a pretty crude measure of a corporate’s culture, when added to the accusations of liberal bias in Google's search engine rankings, and the fact that Eric Schmidt, Google’s Executive Chairman, was closely involved in the strategic planning of Clinton’s campaign, it is understandable why Google may have gained a reputation as a distinctly left-leaning organisation.
Is this a problem?
Aside from the obvious concerns (keenly sharpened by the US election) about the trustworthiness of our news sources, there are important questions about innovation and diversity of thought.
Diversity can strengthen innovation. In many circumstances, a group reaches a better collective decision when its members are more diverse.
Research suggests that staff diversity may increase a firm’s patenting activity, improve its tendency towards radical innovation and increase total income from innovation. Among startups, there is evidence that diverse teams out-perform single-gender teams, on various measures.
However, the evidence about what types of diversity promote innovation, in what circumstances, is actually quite complicated.
Firstly, it should be noted that the story is not all positive: some research actually finds that diverse organisations may suffer from poorer communication and weaker trust, with the net effect of hindering innovation.
Where innovation is enhanced, however, some types of diversity seem to matter more than others. For instance, at least one study found that age diversity was irrelevant to innovation. The same also concluded that, while gender diversity had a positive effect on innovation, it was less significant than differences in industry background or country of origin.
Furthermore, the benefits seem to be more apparent in some types of company than others: factors such as organisational size, complexity of operations, and seniority of staff seem to influence whether or not there are innovation gains from gender diversity.
Importantly, researchers such as Scott E Page suggest that the visible elements of diversity - identity characteristics such as race, gender, sexual orientation, age or religion - are really proxies for (and often determinants of) the core thing that actually matters: cognitive diversity or viewpoint diversity.
According to researchers like Page and others, what is important in many circumstances is that members of a group think differently, leading them to bring different perspectives, try a wider variety of approaches, test a wider range of potential solutions and, in some cases, cancel each other’s biases more effectively.
Returning to Google, even if the firm’s employees simply reflect typical San Francisco voters, rather than political donations, then conservatives are highly likely to be a substantially smaller minority (<15 per cent) than Asian employees (35per cent) or female employees (31 per cent). This means that a whole lot of viewpoint diversity is simply absent.
Fortunately, there are signs that some in Silicon Valley are making efforts to address the problem. Shortly after last year’s Senate enquiry, Facebook COO Sheryl Sandberg reaffirmed the company’s commitment to viewpoint diversity and said in an interview that the company would start including scenarios relating to political bias in the diversity training for its editorial team, as well as racial bias, age bias and gender bias.
Will Google follow suit?
It is difficult to say. But if their CEO, Sundar Pichai, really means what he says in claiming that ‘a diverse mix of voices leads to better discussions, decisions, and outcomes for everyone’, then perhaps, as its next diversity initiative, Google could try introducing an anonymous poll of its staff, specifically aiming to measure viewpoint diversity.