Discover more from Evening Post and Mail
Have a quick chat
On college essay cheaters, chatbots, and the cybersecurity threat we're not even close to being ready to combat
The release of a powerful artificial intelligence chat bot has a lot of technological enthusiasts and observers at high attention, as well it should. The quality of the processing and its ability to generate highly serviceable plain-English responses to human requests is enough to have garnered more than a little attention from some of those who earn a living by writing.
■ In some quarters, it has been noted that continued development in this area will make it impossible for instructors to give their students take-home essays or assignments and expect to gain a fair assessment of student progress. If an AI can answer an essay question and do it passably well, then what's to stop students from letting the computers do all the work?
■ While it would be unwise to trivialize the academic impact, thoughtful observers ought to be fifty times more concerned about what this means for phishing attacks -- most especially for spearphishing. While people have generally adapted to ignoring spam, phishing attacks (scams that depend upon impersonation and misidentification) still generate at least some success for their crooked creators simply through volume. Someone, somewhere, eventually falls for a scam.
■ Spearphishing takes ordinary phishing attacks and makes them more sophisticated by targeting specific users with specific information. This is far more time-intensive than a general-purpose phishing attack, but it raises the odds of success for the attacker. The problem introduced by these much-improved artificial intelligence tools is that they can merge some of the high volumes of classic phishing with the targeting of spearphishing -- which, in combination, makes them much more likely to extract money from unwitting victims.
■ Society is dangerously naive about what's on our cybersecurity doorstep. For a long time, we have been able to depend largely upon hardware and software to offer some protection -- antivirus programs, attachment scanning, and spam filters all do a lot of work to cast a protective bubble over us. But these tools aren't going to be able to keep up with natural-language artificial intelligence in the hands of clever crooks with the incentive to keep hacking human beings. We are rapidly moving from an era of computer vulnerability to one of pure human vulnerability.
■ It's not the only public policy priority out there today, but if we don't amplify consideration and discussion of cybersecurity as a matter for universal responsibility and engagement, we're going to be in deep trouble.