The author argues that AI advancements making software cheaper and faster to fully (and by fully the author means completely patched up, no vulnerabilities) secure will actually result in some negative consequences for software security and privacy.
He begins with a history of encryption and security – the rise of iPhone and end-to-end encryption, which resulted in American national intelligence agencies “going dark”. This led to a legal dispute between Apple and the FBI where the FBI wanted Apple to give hem access to private messages, but Apple refused. The workaround was getting another private entity to simply hack Apple for the FBI.
The episode above shows the ongoing tension between ever more secure software and national intelligence, where software improvements have increased the workload of these agencies. But AI is unprecedented in possibly allowing completely secure software in the future (or so the author believes), which is troublesome for national intelligence agencies because they could be completely locked out of all personal devices no matter how hard they try to break into one. The author believes it’s quite possible that the American government may force companies to start putting intentional backdoors in their systems to make them accessible by intelligence agencies. This has a few negative consequences:
-Non-US governments might want to remove their dependence on US software
-Weakens US’s own systems which will allow opportunities for foreign adversaries to exploit vulnerabilities
1 Comment
The author argues that AI advancements making software cheaper and faster to fully (and by fully the author means completely patched up, no vulnerabilities) secure will actually result in some negative consequences for software security and privacy.
He begins with a history of encryption and security – the rise of iPhone and end-to-end encryption, which resulted in American national intelligence agencies “going dark”. This led to a legal dispute between Apple and the FBI where the FBI wanted Apple to give hem access to private messages, but Apple refused. The workaround was getting another private entity to simply hack Apple for the FBI.
The episode above shows the ongoing tension between ever more secure software and national intelligence, where software improvements have increased the workload of these agencies. But AI is unprecedented in possibly allowing completely secure software in the future (or so the author believes), which is troublesome for national intelligence agencies because they could be completely locked out of all personal devices no matter how hard they try to break into one. The author believes it’s quite possible that the American government may force companies to start putting intentional backdoors in their systems to make them accessible by intelligence agencies. This has a few negative consequences:
-Non-US governments might want to remove their dependence on US software
-Weakens US’s own systems which will allow opportunities for foreign adversaries to exploit vulnerabilities