Web creator Tim Berners-Lee on the future of data

Bil­lions of us use the World Wide Web as our pri­ma­ry tool to inter­act online. Today, its cre­ator Sir Tim Bern­ers-Lee is on a new mis­sion: to ensure data is used appro­pri­ate­ly to cre­ate the pub­lic sec­tor of the future.

Bern­ers-Lee part­nered with arti­fi­cial intel­li­gence (AI) expert Sir Nigel Shad­bolt in 2012 to found the Open Data Insti­tute (ODI). At the ODI Sum­mit in ear­ly Novem­ber the pair of com­put­er sci­en­tists warned that now is a piv­otal moment. As we hur­tle into the dig­i­tal era pow­ered by data-hun­gry algo­rithms and AI, it’s crit­i­cal to col­lab­o­rate with good inten­tions and max­imise the poten­tial of tech­nol­o­gy, for the sake of the plan­et and its inhab­i­tants. 

The accel­er­a­tion of dig­i­tal trans­for­ma­tion neces­si­tat­ed by the coro­n­avirus chaos is excit­ing, but there’s a respon­si­bil­i­ty for author­i­ties around the world to keep pace with this incred­i­ble change. Those in pow­er must set stan­dards, encour­age data to be opened and shared respon­si­bly, and nar­row the ever-widen­ing skills gap. The quick­er that data lit­er­a­cy in both pri­vate and pub­lic sec­tors can be improved, the bet­ter for every­one.

As Bern­ers-Lee points out, the pan­dem­ic has uncon­scious­ly boost­ed pub­lic aware­ness of how data can save and enrich lives. “Some­thing that took off huge­ly was com­mu­ni­ca­tion through data, with the gov­ern­ment telling us to ‘flat­ten the curve’ [and lim­it the spread of the virus],” he says. “I would imag­ine that the data lit­er­a­cy of the gen­er­al pop­u­la­tion has gone up a chunk.”  

Driving change 

By improv­ing their data lit­er­a­cy, lead­ers and mem­bers of the pub­lic could under­stand and chal­lenge how data is pre­sent­ed, Shad­bolt sug­gests. As pub­lic sec­tor tech­nol­o­gy and its appli­ca­tion devel­ops in the com­ing years, fuelled by more and bet­ter-qual­i­ty data, greater scruti­ny will help shape prod­ucts and ser­vices for the dig­i­tal era. 

The open­ing of more data sources will super-charge the pub­lic sec­tor of the future and dri­ve inno­va­tion, says Shad­bolt. The chair of the ODI – who’s been prin­ci­pal of Jesus Col­lege at Oxford Uni­ver­si­ty since 2015, among oth­er roles – points to the suc­cess of open data pio­neer Trans­port for Lon­don (TfL). Often held up as an exem­plar of open data, TfL offers data feeds and guide­lines about air qual­i­ty, cycling, walk­ing, plan­ning and more. 

In 2017, Deloitte cal­cu­lat­ed that TfL’s release of open data gen­er­at­ed annu­al eco­nom­ic ben­e­fits and sav­ings of up to £130 mil­lion for trav­ellers, the cap­i­tal and the organ­i­sa­tion itself. Addi­tion­al­ly, many pri­vate busi­ness­es have tak­en advan­tage and cashed in on the open appli­ca­tion pro­gram­ming inter­faces (APIs).

“Imag­ine that a lot of data rel­e­vant to every­thing cli­mate-relat­ed was just being rou­tine­ly pub­lished using stan­dard APIs,” Shad­bolt con­tin­ues. “It’s what we saw hap­pen with TfL. And there’s just a bunch of sec­tors and areas to go for.”

How­ev­er, it can be dan­ger­ous to blind­ly fol­low data. Shad­bolt won­ders whether Boris Johnson’s refrain dur­ing the pan­dem­ic that the gov­ern­ment would “fol­low the data” to jus­ti­fy its pan­dem­ic-relat­ed deci­sions coro­n­avirus sent out the wrong mes­sage. “It was quite a bad phrase, in some respects,” he says, “because while there should be a basic abil­i­ty to under­stand the data, we need to inter­ro­gate and cri­tique that data.”

Data can be good, but it nev­er gives a com­plete pic­ture

Ques­tion­ing data sources is not just essen­tial to fight fake news on social media and else­where – it will also enable pub­lic sec­tor organ­i­sa­tions to build greater trust, Bern­ers-Lee says. With more con­nect­ed data, they could trig­ger a shift from reac­tive to proac­tive ser­vices. 

It’s a vir­tu­ous cir­cle, because trust­ed and qual­i­ty datasets will widen the pos­si­bil­i­ties and reach of pub­lic sec­tor tech­nol­o­gy and empow­er cit­i­zens. “Prove­nance is impor­tant for data qual­i­ty, and prove­nance is impor­tant for trust,” he says.

Building trust

For exam­ple, Bern­ers-Lee says a doc­tor should be able to look at the dig­i­tal notes of a per­son with dia­betes and open a data nar­ra­tive explain­ing how this diag­no­sis was made and oth­er rel­e­vant his­to­ry. Pub­lic trust in the data used by the pub­lic sec­tor is cen­tral to the adop­tion of tech­nolo­gies and ser­vices, he points out.

The gen­er­al pub­lic seemed to go into dif­fer­ent cat­e­gories regard­ing coro­n­avirus data, Bern­ers-Lee says. Some accept­ed rec­om­men­da­tions for push­ing the curve down, but oth­ers “don’t lis­ten to the same peo­ple as we might. Instead, they find groups of peo­ple –

the con­spir­a­cy the­o­rists – usu­al­ly on social media, who make up all kinds of strange things about the pan­dem­ic, vac­cines or cli­mate change, for that mat­ter.”

Shad­bolt says experts act in good faith with the infor­ma­tion avail­able at a spe­cif­ic time, but their vis­i­bil­i­ty is lim­it­ed if they have scant amounts of data. The wider the vari­ety of good qual­i­ty data sources, the fuller the pic­ture. “We’ve talked a lot about how it’s impor­tant, par­tic­u­lar­ly dur­ing the pan­dem­ic, not to regard the sci­en­tists, medics and peo­ple in white coats as telling you the whole truth,” he says. “They’re try­ing to give the best infor­ma­tion, very often under con­di­tions of con­sid­er­able uncer­tain­ty.” We must take a nuanced approach, he argues, under­stand­ing that “the data can be good, but it nev­er gives a com­plete pic­ture.”

Those in the pub­lic sec­tor and beyond must be “crit­i­cal­ly reflec­tive” of data. “All our respons­es are made, in a sense, stand­ing on the edge of error. But that’s what sci­ence is: it can believe some­thing is wrong and can revise what we believe as these things unfold.” 

While the col­lab­o­ra­tive use of data will cre­ate smarter pub­lic ser­vices in the UK, this approach is cru­cial on a larg­er scale if human­i­ty is to over­come its biggest chal­lenges. It’s been vital in the response to coro­n­avirus, while a coop­er­a­tive, non-com­pet­i­tive and can-do atti­tude is also essen­tial to reduce glob­al warm­ing.

“We’ve just been liv­ing through an exis­ten­tial cri­sis – a glob­al pan­dem­ic – and we’re in the midst of anoth­er one unfold­ing, with the cli­mate chal­lenge,” says Shad­bolt. “Data will be an essen­tial part of [solv­ing this]: the infra­struc­ture, the insti­tu­tions we might need, the trust we have [in its use], and our lit­er­a­cy.”

Sir Patrick Val­lance, the UK’s chief sci­en­tif­ic advi­sor, echoed this view at the 2021 Unit­ed Nations Cli­mate Change Con­fer­ence (COP26). He warned that the chal­lenge of glob­al warm­ing is a greater risk than Covid-19 and more peo­ple will die from it than the pan­dem­ic if the pub­lic sec­tor doesn’t act quick­ly. Val­lance also said the cli­mate cri­sis could last 100 years and require “a com­bi­na­tion of tech­nol­o­gy and behav­iour­al change”.

Prove­nance is impor­tant for data qual­i­ty, and prove­nance is impor­tant for trust

Shad­bolt con­curs but stress­es that open­ing data and boost­ing cross-sec­tor col­lab­o­ra­tion will accel­er­ate mean­ing­ful change on a macro and micro scale and increase the capa­bil­i­ties of pub­lic sec­tor tech­nol­o­gy. “While envi­ron­ment data is in the news because of COP26, there is oth­er infor­ma­tion that can help spur action,” he says, hint­ing that greater trans­paren­cy from pub­lic sec­tor organ­i­sa­tions will ratch­et up pres­sure on pri­vate com­pa­nies to keep clean. For exam­ple, he notes that data on util­i­ty com­pa­nies dis­charg­ing sewage will help the Envi­ron­ment Agency, which strug­gles with funds and sup­port. 

“We are start­ing to gain a sense of what data’s going to make a dif­fer­ence – every­thing from emis­sions to insu­la­tion. There’s a whole net­work of inter­con­nect­ed data types that we can bring togeth­er, much of it held in the pub­lic sec­tor, and some of it held in the pri­vate sec­tor,” he says. “We need to begin that work on joint pub­lic-pri­vate enter­pris­es, though we are begin­ning to see the pri­vate sec­tor, with its com­mit­ments to ESG, say­ing ‘we now have to have a pub­lic pur­pose as well as a pri­vate one.’” Pub­lish­ing some of this data “would be a great first step”, he adds. 

Information advantage

Bern­ers-Lee and Shad­bolt were appoint­ed as infor­ma­tion advi­sors to the gov­ern­ment in June 2009. The duo led the team that devel­oped data.gov.uk, a sin­gle point of access for UK non-per­son­al gov­ern­men­tal pub­lic data. This offers real-time infor­ma­tion on a range of top­ics, such as gov­ern­ment spend­ing, dig­i­tal ser­vice per­for­mance, crime and jus­tice, trans­port and more.

When the pair found­ed the not-for-prof­it ODI nine years ago, the mis­sion was to “con­nect, equip and inspire peo­ple around the world to inno­vate with data”. Almost a decade lat­er, the ODI con­tin­ues to pro­vide free and paid-for train­ing cours­es and learn­ing mate­ri­als both in-house and online. These cov­er the­o­ry and prac­tice sur­round­ing data pub­lish­ing and use. The ODI has long cham­pi­oned open data as a pub­lic good, but always empha­sised that effec­tive gov­er­nance mod­els are nec­es­sary to pro­tect cit­i­zens.

Some 20 months since the start of the coro­n­avirus cri­sis, peo­ple are begin­ning to appre­ci­ate the ODI’s work and con­cerns around data stan­dards. “When the pan­dem­ic began we pro­vid­ed a data pub­li­ca­tion tem­plate,” says Shad­bolt. “The chal­lenge was so many peo­ple want­ed to con­tribute data. It need­ed sort­ing and we had to deter­mine what was help­ful. If there was just a lit­tle more aware­ness around open stan­dards to pub­lish data, so that it is in a more inter­op­er­a­ble for­mat, it would be bet­ter for every­one.”

For pub­lic sec­tor tech­nol­o­gy to thrive, how­ev­er, pub­lic trust is crit­i­cal, says Bern­ers-Lee, who notes a dif­fer­ence in atti­tudes to tech in the UK com­pared to the US. “Typ­i­cal­ly in the UK peo­ple trust the gov­ern­ment and don’t trust [the tech] indus­try, and in the US peo­ple trust indus­try and don’t trust the gov­ern­ment,” he says. More should be done to assuage fears about how tech giants han­dle user data, he adds. “To an extent, it’s how peo­ple are brought up and there­fore cul­tur­al. But for peo­ple in the UK to trust these large Amer­i­can com­pa­nies then you need to have seri­ous leg­is­la­tion and reg­u­la­tion.”

The back­lash against the alleged­ly avari­cious Face­book, which accord­ing to a recent whistle­blow­er puts user engage­ment ahead of safe­ty, is a cau­tion­ary tale for pub­lic sec­tor organ­i­sa­tions seek­ing to embrace tech­nol­o­gy solu­tions and part­ner with com­pa­nies with­out ful­ly know­ing their poli­cies on data pri­va­cy and oth­er ques­tion­able val­ues, sug­gests Bern­ers-Lee. More than ever, at the out­set, dig­i­tal prod­ucts must be “good by design”.

Data man­age­ment is inte­gral to these process­es. Here too the coro­n­avirus has proven use­ful, test­ing the robust­ness of so-called ‘trust­ed research envi­ron­ments’. “In these envi­ron­ments, the data stays behind a fire­wall and it’s mod­elled and analysed with tools that can go behind the fire­wall,” Shad­bolt explains. “The data nev­er actu­al­ly leaves the high­ly secure data stor­age areas where 47 mil­lion patient records are linked, but incred­i­ble insights are gained.”

Offer­ing an alter­na­tive, he says: “The oth­er solu­tion is to leave the data with the peo­ple who gen­er­ate it, which is very local. There are dif­fer­ent tech­ni­cal solu­tions there and there are dif­fer­ent insti­tu­tions we can build to share this. It’s a com­pli­cat­ed area, but the ODI is look­ing very care­ful­ly at mak­ing data shar­ing more effec­tive.”

Unfinished business

What does the future hold for the ODI as it nears its 10-year anniver­sary? “We start­ed off explain­ing to peo­ple work­ing in the pub­lic sec­tor how to put your data on the web,” says Bern­ers-Lee. Now, how­ev­er, “we realise it’s impor­tant to cov­er the whole spec­trum, from pub­lic to pri­vate, but it’s also about devel­op­ing poli­cies as well.”

This assess­ment chimes with Shad­bolt. “There is unfin­ished busi­ness,” he says. “The whole com­mit­ment to get­ting data out there was start­ed with open data ini­tia­tives that were very much focused around the pub­lic sec­tor – every­thing from hos­pi­tal data to edu­ca­tion­al data to trans­port data. That work has gone well. We’re now look­ing at extend­ing those learn­ings. As gov­ern­ments move on [in their dig­i­tal trans­for­ma­tion jour­neys], you want to ensure that momen­tum is kept up and that the infra­struc­ture is there to help sus­tain pub­lish­ing the data out.”

Return­ing to the glob­al cli­mate cri­sis, he says of the ODI’s mis­sion: “We did antic­i­pate that in try­ing to build a trust­ed research data ecosys­tem it would become one of the con­se­quen­tial ques­tions for the future of the plan­et and the future of our well­be­ing. There’s a huge amount of work to do. We’re try­ing to make sense of it in terms of pro­grammes of work, from data lit­er­a­cy to insti­tu­tions, from ethics to infra­struc­ture.”

Shad­bolt adds: “Fun­da­men­tal­ly the ODI’s work is about lis­ten­ing, it’s about try­ing to take ideas and put them in a for­mat that allows that to scale. We may be an organ­i­sa­tion of 60-odd peo­ple but we think we can have a fan­tas­tic impact and so we need to reach out and sus­tain our­selves to make a bet­ter future.”

---