6 ways AI can help save the planet

1. Conserving species

The Liv­ing Plan­et Index pro­duced by WWF esti­mates that wildlife pop­u­la­tion sizes have dropped by 68 per cent since 1970. The char­i­ty advo­cates the use of arti­fi­cial intel­li­gence (AI) as a tool of con­ser­va­tion tech­nol­o­gy to mon­i­tor and curb this alarm­ing rate of decline.

One of the most use­ful appli­ca­tions is in acoustic mon­i­tor­ing, record­ing the sounds of wildlife ecosys­tems on weath­er­proof sen­sors. Many ani­mals, from birds and bats to mam­mals and even inver­te­brates, use sound for com­mu­ni­ca­tion, nav­i­ga­tion and ter­ri­to­r­i­al defence, pro­vid­ing reams of rich data on how a species pop­u­la­tion is doing. AI pro­vides a fast and cost-effec­tive way to analyse hours of record­ings for pat­terns of behav­iour.

Con­ser­va­tion Met­rics, a Cal­i­for­nia-based com­pa­ny, has used acoustic lis­ten­ing and machine-learn­ing to mon­i­tor endan­gered pop­u­la­tions of both red-legged frogs in San­ta Cruz, divert­ing water to help them mate suc­cess­ful­ly, and the for­est ele­phants of the Cen­tral African Repub­lic, help­ing to pro­tect them from poach­ers.

Facial recog­ni­tion tech­nol­o­gy is anoth­er appli­ca­tion of AI that could help track wildlife pop­u­la­tions, when com­bined with cam­era traps in the wild. BearID, an open-source appli­ca­tion, which was trained on brown bears in Cana­da and the Unit­ed States, is a recent AI tri­umph as, unlike pri­mates, zebras or giraffes, bears don’t have dis­tin­guish­ing fea­tures, so the deep-learn­ing algo­rithm had to find pat­terns in their facial make-up instead. The researchers hope this AI will be used to mon­i­tor oth­er species in the future.

2. Improving recycling

More than 2.1 bil­lion tonnes of rub­bish is gen­er­at­ed in the world each year, yet only 16 per cent of it is recy­cled, accord­ing to research by Maple­croft. To make mat­ters worse, a quar­ter of waste put into the recy­cling is not actu­al­ly recy­clable at all, hin­der­ing the whole process.

Sev­er­al star­tups are now look­ing at how AI and sus­tain­abil­i­ty goals can be com­bined to make recy­cling more effi­cient, even when deal­ing with mixed mate­ri­als. Col­orado-based AMP Robot­ics uses an AI-pow­ered robot with opti­cal sen­sors to quick­ly iden­ti­fy rub­bish as it pass­es on a con­vey­or belt. It then sorts it with its robot­ic arms, using the company’s AMP Neu­ron AI plat­form, which can recog­nise dif­fer­ent tex­tures, colours, shapes, sizes and even brand labels.

The AI con­stant­ly updates itself and is designed to run 24/7. It has already been rolled out in the Unit­ed States, Cana­da and Japan, and will soon be com­ing to Europe.

In Bali, Gring­go Tech has designed an image recog­ni­tion tool to help infor­mal waste col­lec­tors iden­ti­fy the dif­fer­ent mon­e­tary val­ues of var­i­ous recy­clable mate­ri­als. In a pilot study, it improved recy­cling rates by 35 per cent. They’re now work­ing with Google to build AI into the plat­form to help improve how quick­ly and effi­cient­ly the sys­tem can cat­e­gorise waste.

3. Protecting forests

Forests are home to 80 per cent of the world’s ter­res­tri­al bio­di­ver­si­ty, and they absorb and store a third of cur­rent car­bon emis­sions. Halt­ing the loss and degra­da­tion of for­est ecosys­tems is essen­tial to meet­ing the objec­tives of the Paris Agree­ment on cli­mate change, accord­ing to the Inter­na­tion­al Union for Con­ser­va­tion of Nature.

Rain­for­est Con­nec­tion seeks to com­bat ille­gal log­ging using acoustic mon­i­tor­ing in forests on hid­den solar-pow­ered smart­phones, which have been recy­cled from con­sumer use. The char­i­ty then uses AI to analyse this sound data in real time. If the AI detects the sounds of chain­saws, log­ging trucks or gun­shots, an alert is sent to rangers. Accord­ing to Rain­for­est Con­nec­tion, research shows that if ille­gal log­gers are inter­rupt­ed once or twice, they leave and don’t return until the next
log­ging sea­son.

Dryad Net­works has secured seed fund­ing to use the inter­net of things and AI to detect wild­fires. Dryad uses AI-based solar-pow­ered sen­sors to cap­ture gas­es emit­ted at the smoul­der­ing stage of a wild­fire which, com­bined with real-time analy­sis of tem­per­a­ture, humid­i­ty, air pres­sure and wind data, will alert for­est rangers when a wild­fire is immi­nent. They are also devel­op­ing a long-range wire­less envi­ron­men­tal mon­i­tor­ing sen­sor net­work to cov­er large for­est areas where there is no mobile-phone sig­nal.

4. Cutting air pollution

Nine in ten of the world’s urban res­i­dents breathe pol­lut­ed air, prompt­ing the Unit­ed Nations to make access to cycling, walk­ing or pub­lic trans­porta­tion one of its 17 Sus­tain­able Devel­op­ment Goals.

To meet this chal­lenge, Lon­don-based Vivac­i­ty uses AI tech­nol­o­gy to cap­ture and clas­si­fy live trans­port usage with the goal of enabling more envi­ron­men­tal­ly sus­tain­able trans­port use in cities. The com­pa­ny has been work­ing with Trans­port for Lon­don since 2018 to deter­mine where new cycling infra­struc­ture should be tar­get­ed.

London’s Walk­ing and Cycling Com­mis­sion­er Dr Will Nor­man says: “By get­ting more peo­ple cycling and walk­ing, we can help to tack­le con­ges­tion and pol­lu­tion in Lon­don and improve our health. Our Healthy Streets approach is based on evi­dence and data, and we wel­come new tech­nol­o­gy that sup­ports this.” 

Vivacity’s AI has allowed local author­i­ties across the UK to assess the effec­tive­ness of their tem­po­rary street lay­outs to encour­age phys­i­cal­ly active trav­el dur­ing the coro­n­avirus cri­sis. The com­pa­ny has also helped Trans­port for Greater Man­ches­ter roll out smart junc­tions across the city, which pri­ori­tise pedes­tri­ans and cyclists over motor-vehi­cle traf­fic.

5. Minimising food waste

Some 9.5 mil­lion tonnes of food is wast­ed in the UK every year, accord­ing to the Waste and Resources Action Pro­gramme, 70 per cent of which could be avoid­ed. The waste, which includes food from super­mar­kets, house­holds and hos­pi­tal­i­ty, gen­er­ates 25 mil­lion tonnes of green­house gas emis­sions.

Win­now is work­ing with HCL Tech­nolo­gies to use AI to tack­le the prob­lem in hos­pi­tal­i­ty, where their data shows up to 15 per cent of pur­chased food is being wast­ed. Win­now Vision is an AI tool that takes pic­tures of food as it’s thrown into the bin, teach­ing itself to recog­nise what’s been dis­card­ed and track­ing the data. IKEA has deployed Win­now Vision in its UK stores, cut­ting food waste by an aver­age of 50 per cent.

Last year, UK super­mar­kets signed up to a gov­ern­ment pledge to halve food waste by 2030. Accord­ing to data from Blue Yon­der, using AI in super­mar­ket sup­ply chains could help the UK’s eight largest retail­ers cut sev­en tonnes of food waste a year, sav­ing £144 mil­lion. As Wayne Sny­der, vice pres­i­dent of retail strat­e­gy, Europe, Mid­dle East and Africa, at Blue Yon­der says: “AI mon­i­tors goods from farm to fork, result­ing in an increased under­stand­ing of the envi­ron­men­tal impacts across the sup­ply chain and iden­ti­fi­ca­tion of the areas that need improv­ing.”

6. Reducing sewage pollution

Raw sewage was dis­charged onto beach­es in the UK almost 3,000 times over the last year, accord­ing to a report by Surfers Against Sewage. The envi­ron­men­tal char­i­ty advo­cates stricter mon­i­tor­ing of sea and riv­er pol­lu­tion, and oper­ates an app called the Safer Seas Ser­vice, which warns swim­mers, surfers and oth­er water users when untreat­ed sewage has been released at their beach.

But the app, which began in 2010 as a text alert sys­tem, relies on vol­un­tary data pro­vid­ed by water com­pa­nies, which isn’t always reli­able. So, this year, Surfers Against Sewage added a health report func­tion to the app, using a cit­i­zen sci­ence approach to warn oth­ers about beach clean­li­ness issues in real time, but also to hold water com­pa­nies to account. South­ern Water, for exam­ple, had released no noti­fi­ca­tions dur­ing 2020 due to report­ing mech­a­nism errors, yet over 20 per cent of health reports sub­mit­ted to Surfers Against Sewage alleged­ly came from beach­es with­in South­ern Water’s juris­dic­tion.

In the future, appli­ca­tion of AI will enable even more pre­cise, live sea­wa­ter qual­i­ty assess­ments. Sci­en­tists work­ing with the Nation­al Research Foun­da­tion of Korea have already shown that arti­fi­cial neur­al net­work mod­els can accu­rate­ly pre­dict micro­bial con­t­a­m­i­na­tion at beach­es, using vari­ables includ­ing tides, tem­per­a­tures, wind speed and direc­tion, rain­fall and recent sewage dis­charges. South­ern Water has set a tar­get of zero pol­lu­tion inci­dents by 2040 and say they will use state-of-the-art machine-learn­ing in that mis­sion.