How fear of los­ing ad­ver­tis­ing dol­lars forced a re­sponse

The Guardian - - News Technology - Alex Hern Tech­nol­ogy ed­i­tor

It took less than two hours for Face­book to re­act – but re­act it did, and with good rea­son. At 5pm on Fri­day, Unilever, one of the world’s largest ad­ver­tis­ers – with a port­fo­lio of prod­ucts that ranges from Mar­mite to Vase­line – sud­denly an­nounced it was pulling all ad­ver­tis­ing from Face­book, In­sta­gram and Twit­ter in the US.

Given the “po­larised at­mos­phere in the US”, Unilever said, and with the sig­nif­i­cant work left to be done “in the ar­eas of di­vi­sive­ness and hate speech … con­tin­u­ing to ad­ver­tise on these plat­forms at this time would not add value to peo­ple and so­ci­ety”.

At 6.47pm, Face­book scram­bled. Mark Zucker­berg, the com­pany said, would be “going live on his Face­book page” to dis­cuss the com­pany’s racial jus­tice work – and 13 min­utes af­ter that, the world’s most pow­er­ful chief ex­ec­u­tive ap­peared on screens.

Ap­pear­ing hum­bled, he an­nounced a se­ries of new poli­cies, in­clud­ing a ban on hate­ful con­tent that tar­gets im­mi­grants, and fur­ther re­stric­tions on posts mak­ing false claims about vot­ing.

Asad Moghal, se­nior dig­i­tal and con­tent man­ager at By­field Con­sul­tancy, said Unilever’s ac­tion was al­ways going to force Zucker­berg to re­spond.

“When such an in­ter­na­tional gi­ant de­cides that in­ac­tion is no longer an op­tion to tackle racist and dis­crim­i­na­tory lan­guage, then the so­cial me­dia busi­nesses need to lis­ten up. By tak­ing fi­nan­cial ac­tion, a com­pany the size of Unilever can ef­fect change and force the hand of Twit­ter and Face­book – the busi­ness has de­cided it needs to pro­tect its brand rep­u­ta­tion and can no longer be as­so­ci­ated with plat­forms that de­liver hate speech and di­vi­sive con­tent. But what will really ef­fect change is if this move cre­ates a domino ef­fect and other big-name cor­po­ra­tions re­move in­vest­ment from the plat­forms.”

The swathe of an­nounce­ments from Zucker­berg marked the first con­ces­sions from Face­book to­wards the aims of a US coali­tion, Stop Hate for Profit, that was formed in the wake of the killing of Ge­orge Floyd in May by Min­neapo­lis po­lice.

But the group’s lead­ers say Face­book’s tweaks do not go far enough, and re­it­er­ated their calls for a month-long global ad­ver­tiser boy­cott start­ing to­mor­row.

The real dan­ger for Face­book is if other brands de­cide they can do with­out the plat­form too.

This crisis has been a long time in the mak­ing – and shows no sign of going away.

Face­book has his­tor­i­cally taken a less cen­so­ri­ous ap­proach with hate speech than it has with other con­tro­ver­sial ar­eas, such as nu­dity – in part out of a be­lief in the in­her­ent am­bi­gu­ity of of­fen­sive speech, and in part due to the dif­fi­culty of au­tomat­ing such work. Iden­ti­fy­ing hate speech is re­liant on knowl­edge of con­text, cus­tom and cul­ture which can be hard to teach hu­man mod­er­a­tors, let alone ma­chines.

In re­cent years Face­book has made strides in the area. In the third quar­ter of 2017, ac­cord­ing to its com­mu­nity stan­dards re­port, the so­cial net­work found just under a quar­ter of in­ci­dences of hate speech

‘Un­like Twit­ter … we be­lieve that if a post in­cites vi­o­lence, it should be re­moved re­gard­less of whether it is news­wor­thy’

by it­self; the other three-quar­ters was only re­moved af­ter site users man­u­ally flagged it to mod­er­a­tors, who then took ac­tion. By this spring, the pro­por­tions had re­versed: 88% of re­moved hate speech was found by Face­book’s own tools, al­low­ing it to re­move or re­strict al­most four times as much ma­te­rial.

But work­ing against Face­book’s tech­ni­cal ex­per­tise was an­other fac­tor: the US pres­i­dent.

As far back as 2015, ac­cord­ing to the Wash­ing­ton Post, Face­book was strug­gling with how to deal with a man who, first as a can­di­date and then as pres­i­dent, pushed the lim­its of what was al­lowed to be posted.

In­stead, Face­book has steadily tweaked its own rules to avoid an­ger­ing Don­ald Trump: in­tro­duc­ing in 2015 an ex­cep­tion for “po­lit­i­cal dis­course” to al­low a video call­ing for a ban on Mus­lims en­ter­ing the US to stay up, for in­stance, or lim­it­ing ef­forts to tackle “false news” out of a fear that do­ing so would dis­pro­por­tion­ately hit right-lean­ing pages and posters.

In the protests af­ter Floyd’s death, Trump again tested the bound­aries, post­ing on Face­book and Twit­ter a mes­sage that “when the loot­ing starts, the shoot­ing starts”.

Twit­ter, not­ing the racist history of the phrase, and in­ter­pret­ing it as a po­ten­tial call for vi­o­lence, re­stricted the tweet, pre­vent­ing it from be­ing replied to or liked, and hid it be­hind a warn­ing declar­ing that it broke its rules. But it left the tweet up, cit­ing the in­her­ent news­wor­thi­ness of a state­ment by an elected of­fi­cial with mil­lions of fol­low­ers.

On Face­book, how­ever, the post was un­touched. In a post on his per­sonal page, Zucker­berg said he in­ter­preted the state­ment not as an in­cite­ment to vi­o­lence but as “a warn­ing about state ac­tion”.

“Un­like Twit­ter,” he wrote, “we do not have a pol­icy of putting a warn­ing in front of posts that may in­cite vi­o­lence be­cause we be­lieve that if a post in­cites vi­o­lence, it should be re­moved re­gard­less of whether it is news­wor­thy, even if it comes from a politi­cian.”

The de­ci­sion be­came a flash­point for lin­ger­ing un­ease about Face­book’s wider is­sues with tack­ling hate – as did Zucker­berg’s de­ci­sion, a week ear­lier, to ap­pear on Fox News to de­fend a dif­fer­ent Trump post, on mail-in vot­ing, say­ing he did not think his com­pany should be­come the “ar­biter of truth”.

Face­book staff be­gan to speak out on so­cial me­dia, hold­ing a vir­tual walk­out to em­pha­sise that “do­ing noth­ing is not ac­cept­able”.

The firm’s pre­car­i­ously em­ployed mod­er­a­tors joined in, risk­ing their con­tracted-out jobs to de­cry the “white ex­cep­tion­al­ity and fur­ther le­git­imi­sa­tion of state brutality”.

Even sci­en­tists funded by Zucker­berg’s per­sonal char­ity, the Chan Zucker­berg Ini­tia­tive, spoke out, call­ing Trump’s post “a clear state­ment of in­cit­ing vi­o­lence”.

Then in May, with some fan­fare, Zucker­berg ap­pointed an over­sight board – a ros­ter of ex­perts that will have the power to over­rule Face­book’s mod­er­a­tion de­ci­sions.

It in­clude Helle Thorn­ingSch­midt, the former prime min­is­ter of Den­mark, the No­bel peace lau­re­ate Tawakkol Kar­man, and Alan Rus­bridger, the former ed­i­tor-in-chief of the Guardian. But the dif­fi­culty of set­ting up a new or­gan­i­sa­tion in the age of coro­n­avirus means that the board was unable to take the heat off Zucker­berg.

“Zucker­berg’s strat­egy of deal­ing with Trump is an in­co­her­ent blend of two lead­er­ship ap­proaches,” said Chris Moos, a lead­er­ship ex­pert and teach­ing fel­low at Ox­ford University’s Saïd busi­ness school.

Where some at­tempt to find “prac­ti­cal ap­proaches for deal­ing with ten­sions” they en­counter at work, and others ap­peal “to high­erorder prin­ci­ples”, Zucker­berg tries both and suc­ceeds at nei­ther, he said. “On the one hand, he has en­gaged a wide set of stake­hold­ers into the de­bate, throw­ing money at ini­tia­tives to build racial jus­tice and voter en­gage­ment. On the other, the Face­book CEO has tried to rise above the con­tro­versy by mak­ing it clear that his com­pany will be erring on the side of free ex­pres­sion, ‘even when it’s speech we strongly and vis­cer­ally dis­agree with’.”

Zucker­berg can never be re­moved from his position. While he owns only 14% of the com­pany, the spe­cial class of shares he holds means he con­trols 57% of the vot­ing rights at board meet­ings. But em­ployee pres­sure can hurt him, pro­fes­sion­ally and per­son­ally: if Face­book no longer seems like a pleas­ant and re­ward­ing work­place, the com­pany will strug­gle to hire and re­tain the highly skilled staff it re­lies on to com­pete in Sil­i­con Val­ley.

In June, the Stop Hate for Profit cam­paign found an­other weak point for the site: ad­ver­tis­ers. While Face­book takes some rev­enue di­rectly from users, the vast ma­jor­ity of the com­pany’s $70.7bn (£57bn) an­nual rev­enue comes from ad­ver­tis­ing.

On 17 June, Color of Change – the or­gan­i­sa­tion be­hind Stop Hate for Profit – launched a pub­lic re­quest: for “all ad­ver­tis­ers to stand in sol­i­dar­ity with Black Face­book users and send the mes­sage to Face­book that they must change their prac­tices by paus­ing all ad­ver­tis­ing on Face­book-owned plat­forms for the month of July 2020.”

Many of those ad­ver­tis­ers were al­ready un­com­fort­able about their spend on Face­book be­fore the lat­est cam­paign. The site, as with all pro­gram­matic ad­ver­tis­ing, can have “brand safety” is­sues when com­pa­nies find their mes­sages next to ex­treme or hate­ful con­tent.

At a macro level, mean­while, mar­keters are all too aware of the risks of help­ing con­sol­i­date the “du­op­oly” of Face­book and Google, who be­tween them have se­cured the ma­jor­ity of the ad­ver­tis­ing in­dus­try’s growth.

But even if the Stop Hate for Profit cam­paign was push­ing at an open door, the suc­cess has been sur­pris­ing. By the end of the first week, Patag­o­nia, North Face and free­lanc­ing plat­form Up­work had signed on. And Unilever’s de­ci­sion on Fri­day to pause ad­ver­tis­ing un­til Novem­ber – al­beit only within the US, and with­out di­rectly cit­ing the cam­paign – opened the flood­gates.

Over the week­end, it was joined by other megabrands, in­clud­ing Coca-Cola and al­co­hol con­glom­er­ate Beam Sun­tory.

“Let’s be hon­est,” said Moghal, “these tech plat­forms have gen­er­ated in­come and in­ter­est from this di­vi­sive con­tent. They won’t change their prac­tices un­til they be­gin to see a sig­nif­i­cant cut to their rev­enue.”

With the boy­cott of­fi­cially start­ing to­mor­row, the cam­paign­ers are not eas­ing up. In fact, suc­cess has only driven higher am­bi­tions.

“The next fron­tier is global pres­sure,” Jim Steyer, the chief ex­ec­u­tive of Com­mon Sense Me­dia, a non­profit or­gan­i­sa­tion, told Reuters yes­ter­day.

While some ad­ver­tis­ers, in­clud­ing North Face and Patag­o­nia, have ex­panded their boy­cotts glob­ally, others are cur­rently con­tent to only with­hold spend­ing in the US.

If even that is enough to get Zucker­berg in front of a cam­era in less than two hours, the cam­paign­ers hope world­wide ac­tion could mo­ti­vate last­ing change.

PHO­TO­GRAPH: PA­TRICK SE­MAN­SKY/AP

Don­ald Trump poses at a Wash­ing­ton church days af­ter post­ing on Face­book that ‘when the loot­ing starts, the shoot­ing starts’

PHO­TO­GRAPH: CHANDAN KHANNA/ AFP VIA GETTY IMAGES

A protest in Min­neapo­lis over Ge­orge Floyd’s death. Stop Hate for Profit was formed af­ter the killing

PHO­TO­GRAPH: JOSH EDELSON/AFP/GETTY IMAGES

▼ Mark Zucker­berg on stage in Cal­i­for­nia. It took Face­book’s CEO two hours to re­act to Unilever’s move

Newspapers in English

Newspapers from UK

© PressReader. All rights reserved.