Is any de­fense pos­si­ble?

The Week (US) - - News 11 -

Tech­nol­o­gists are work­ing fu­ri­ously on new strate­gies. One idea is to use ar­ti­fi­cial in­tel­li­gence to mon­i­tor net­works for sus­pi­cious ac­tiv­ity that would oth­er­wise go un­no­ticed, act­ing as a dig­i­tal “im­mune sys­tem.” Some re­searchers are de­vel­op­ing hard­ware that’s built for se­cu­rity from the ground up, in­clud­ing com­puter chips that can’t be fooled by bo­gus in­struc­tions. But even un­der ideal con­di­tions, the na­ture of com­put­ing makes at­tacks in­evitable. It’s es­ti­mated that pro­gram­mers com­mit about 50 er­rors per every 1,000 lines of code. The lat­est ver­sion of Win­dows is roughly 50 mil­lion lines long, and the An­droid smart­phone op­er­at­ing sys­tem has 12 mil­lion lines of code. Even after rig­or­ous check­ing, bugs get through. Then there’s the po­ten­tial for hu­man er­ror: Hil­lary Clin­ton’s cam­paign chairman, John Podesta, made the 2016 Rus­sian hack pos­si­ble by click­ing on a spearphish­ing link, giv­ing hack­ers ac­cess to his emails. “The at­tack­ers only have to find one weak­ness,” says Kath­leen Fisher, a com­puter sci­en­tist at Tufts Univer­sity. “The de­fend­ers have to plug every sin­gle hole, in­clud­ing ones they don’t know about.”

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.