Is the UK still an AI leader?

The UK has been at the cut­ting edge of arti­fi­cial intel­li­gence (AI) inno­va­tion, from Alan Tur­ing, the pio­neer­ing math­e­mati­cian and com­put­er vision­ary, who launched the field, to DeepMind’s Alpha­Go, the first com­put­er pro­gram to defeat a pro­fes­sion­al Go play­er in 2015.

Sev­er­al pio­neer­ing AI com­pa­nies were found­ed in the UK, includ­ing Deep­Mind, SwiftKey and Mag­ic Pony, all of which were acquired by US com­pa­nies – Google, Microsoft and Twit­ter – for $500 mil­lion, $250 mil­lion and $150 mil­lion, respec­tive­ly. Over the last few years, the UK gov­ern­ment has launched its Office for AI and Cen­tre for Data Ethics and Inno­va­tion. But is the UK still an AI leader?

In 2019, McK­in­sey Glob­al Insti­tute placed the UK in the top quar­tile for “AI readi­ness”. How is the UK main­tain­ing this posi­tion in a com­pet­i­tive land­scape, both in a busi­ness sense and a gov­ern­men­tal one?

No coun­try can hold a can­dle to the Unit­ed States and Chi­na when it comes to AI, but the UK is one of Europe’s lead­ers, accord­ing to the McK­in­sey report. The UK is glob­al­ly in the top quar­tile for research, start­up invest­ment, dig­i­tal absorp­tion, inno­va­tion foun­da­tion and ICT con­nect­ed­ness. It does, how­ev­er, rank low­er on automa­tion poten­tial and human cap­i­tal.

Bringing AI skills from the top to a wider population 

The UK has many lead­ing researchers, who are pub­lished in the top aca­d­e­m­ic jour­nals. Chris­tine Fos­ter, chief com­mer­cial offi­cer at The Alan Tur­ing Insti­tute, says: “The UK has emi­nent researchers, such as Christi­na Pagel, who works on math­e­mat­i­cal tools to sup­port deliv­ery of health ser­vices; Mark Giro­la­mi, who is devel­op­ing and apply­ing advanced sta­tis­ti­cal and com­pu­ta­tion­al tech­niques to engi­neer­ing chal­lenges; Max­ine Mack­in­tosh, who has found­ed One HealthTech,” which sup­ports under-rep­re­sent­ed groups in health tech inno­va­tion. There are many oth­ers.

Lee Har­land, founder and chief sci­en­tif­ic offi­cer at SciB­ite, an Else­vi­er com­pa­ny, says: “We’re very good at the basic sci­ence; a strength of the UK has always been our intel­lec­tu­al out­put. The Cam­bridge-Oxford-Lon­don tri­an­gle is a hub for tal­ent. Because AI is a broad skill that can fit just as much into gam­ing as it does into health­care, with­in the tri­an­gle there is a lot of oppor­tu­ni­ty for peo­ple to move around, even into dif­fer­ent indus­tries, with­out trekking halfway across the world.”

The ques­tion is whether this skillset fil­ters down to a broad­er pop­u­la­tion. “Recruit­ing tal­ent from out­side the UK will always be impor­tant, but we need to bring the AI skills clos­er to our schools and uni­ver­si­ties,” says Har­land.

“You don’t need a degree in mechan­i­cal engi­neer­ing to dri­ve a car and you don’t need a degree in sta­tis­tics to use machine-learn­ing. There are some great ini­tia­tives for data sci­ence and AI-cen­tric cours­es appear­ing in our uni­ver­si­ties; this needs to be accel­er­at­ed and cas­cad­ed down, at least con­cep­tu­al­ly, to school age.” Like most coun­tries, the UK faces a short­age of peo­ple with advanced tech­no­log­i­cal skills. Wider edu­ca­tion could rem­e­dy that.

Turning innovation into scalable business opportunities

Fos­ter adds: “In the UK, we are flu­id­ly con­nect­ing and con­ven­ing across the pub­lic sec­tor, pri­vate sec­tor and third sec­tor. Look at the AI Coun­cil [an inde­pen­dent expert com­mit­tee that advis­es the gov­ern­ment]; it’s a great exam­ple of what can hap­pen when peo­ple from indus­try, pub­lic sec­tor and acad­e­mia come togeth­er, shar­ing their broad range of back­ground and exper­tise to the AI ecosys­tem.”

 Despite slight­ly high­er invest­ment in AI, the UK lags behind France, Ger­many, Japan and South Korea when it comes to AI patents, accord­ing to McK­in­sey. What’s more, an inde­pen­dent review com­mis­sioned by the gov­ern­ment not­ed that “uni­ver­si­ties should pro­mote stan­dard­i­s­a­tion in trans­fer of intel­lec­tu­al prop­er­ty”. This would make it eas­i­er to cre­ate spin-out busi­ness­es.

Tak­ing an idea and turn­ing it into a busi­ness takes a com­bi­na­tion of fac­tors, says Har­land. “There are a lot of insti­tu­tions out there to advise – Inno­vate UK, Dig­i­tal Cat­a­pult – but it’s often very obtuse in terms of what they can do and how they help.” He says oth­er Euro­pean coun­tries are bet­ter at being explic­it about which agen­cies do what for star­tups. “I think it’s very hard to under­stand that in the UK land­scape,” says Har­land.

A long-term strategy for Britain to be an AI leader

There is some­thing of a “space race” in the AI realm, says Dr Michael Feindt, strate­gic advis­er of Blue Yon­der. Amer­i­ca is invest­ing fifty times more in AI than the UK, and Chi­na is invest­ing eight times more. “We are increas­ing­ly see­ing promis­ing UK star­tups being acquired by large US com­pa­nies before they can mature, lim­it­ing the UK’s abil­i­ty to make up ground on oth­er coun­tries,” says Feindt.

His­tor­i­cal­ly, many inno­va­tions in the com­put­er indus­try have been pio­neered by women. The first com­put­er pro­gram­mer was Lady Ada Lovelace, while actress Hedy Lamarr invent­ed the tech­nol­o­gy that enabled wifi, GPS and Blue­tooth. In the mid-80s, almost 40 per cent of US com­put­er grad­u­ates were women. But the AI indus­try now faces what Bill Gates called the “sea of dudes prob­lem”. A greater diver­si­ty of peo­ple and data would coun­ter­act some of the bias that algo­rithms have ingest­ed so far.

To stay at the fore­front of AI, the UK needs a long-term strat­e­gy span­ning ten to fif­teen years, rather than just one or three, argues Fos­ter at The Alan Tur­ing Insti­tute. This strat­e­gy needs to ensure data is more acces­si­ble to AI com­pa­nies, that inno­v­a­tive pilots can be scaled and eth­i­cal frame­works applied.

“We have a long his­to­ry in AI. Our researchers know they’re stand­ing on the shoul­ders of giants and that we have the abil­i­ty to move the whole field for­ward,” she con­cludes.