‘Bur­glar bots’ will slip in the let­ter box

Irish Independent - - World News - Martin Evans

FOR many home­own­ers, the traditiona­l way to pro­tect against bur­glars has been to en­sure doors and win­dows are locked and per­haps leave a light on.

But such ges­tures might prove fu­tile in years to come af­ter sci­en­tists warned that the next gen­er­a­tion of home in­vaders could be ro­bots that are pro­grammed to gain en­try through cat flaps or let­ter boxes.

Us­ing Ar­ti­fi­cial In­tel­li­gence (AI), small ro­bots are be­ing de­vel­oped that could breach traditiona­l se­cu­rity safe­guards. De­liv­ered through small open­ings such as cat flaps, they could scan a per­son’s home in or­der to re­trieve keys that could then al­low a hu­man bur­glar to en­ter.

Al­ter­na­tively, sci­en­tists be­lieve more ad­vanced de­vices could use AI to search a prop­erty them­selves for valu­ables, or cash, us­ing cam­eras to scan and ac­cess dif­fer­ent rooms.

The ro­bots could also be used to de­ter­mine whether any­body is at home, re­lay­ing the in­for­ma­tion to a hu­man op­er­a­tor who could then break in if the coast is clear.

The fright­en­ing prospect is just one area in which sci­en­tists and po­lice be­lieve AI could be used by crim­i­nals to ex­ploit peo­ple in the fu­ture.

A study, pub­lished in ‘Crime Sci­ence’ by re­searchers at Lon­don’s UCL, iden­ti­fied a range of crim­i­nal op­por­tu­ni­ties that tech­no­log­i­cal ad­vances could cre­ate. While the use of so-called “bur­glar bots” is re­garded as a low-re­ward crime, sci­en­tists and po­lice are con­cerned about “deep­fake” videos and im­ages that could ex­ploit un­sus­pect­ing vic­tims.

Us­ing so­phis­ti­cated soft­ware, crim­i­nals are able to gen­er­ate im­per­son­ations of peo­ple, which could be used to per­suade peo­ple to part with money or se­cure pass­words.

Po­lice fear un­scrupu­lous crim­i­nal gangs could gen­er­ate a video of some­one from ma­te­rial freely avail­able on­line and use it to per­suade their el­derly par­ents to send them money.

The re­searchers also high­lighted the po­ten­tial risks posed by the roll-out of driver­less cars, which they warned could be used by ex­trem­ists to carry out ter­ror at­tacks.

Pro­fes­sor Lewis Grif­fin, from UCL’s com­puter sci­ence depart­ment, the se­nior au­thor of the re­port, said: “As the ca­pa­bil­i­ties of AI-based tech­nolo­gies ex­pand, so too has their po­ten­tial for crim­i­nal ex­ploita­tion. To ad­e­quately pre­pare for pos­si­ble AI threats, we need to iden­tify what these threats might be, and how they may im­pact our lives.”

Newspapers in English

Newspapers from Ireland

© PressReader. All rights reserved.