Am I getting too old to spot an AI fake?

𝐀𝐦 𝐈 𝐠𝐞𝐭𝐭𝐢𝐧𝐠 𝐭𝐨𝐨 𝐨𝐥𝐝 𝐭𝐨 𝐬𝐩𝐨𝐭 𝐚𝐧 𝐀𝐈 𝐟𝐚𝐤𝐞? My kids regularly tease me about AI-generated music and videos I don’t recognise. Sometimes they’re right. Sometimes I’m not so sure.

So I decided to put their confidence to the test: I played them a song from 2015—which, to be fair, does sound suspiciously AI-generated—and asked whether it was AI. (𝐿𝑖𝑛𝑘 𝑖𝑛 𝑡ℎ𝑒 𝑐𝑜𝑚𝑚𝑒𝑛𝑡𝑠. 𝑌𝑜𝑢 𝑡𝑒𝑙𝑙 𝑚𝑒: 𝑑𝑖𝑑 𝐼 𝑝𝑙𝑎𝑦 𝑡ℎ𝑒 𝑠𝑜𝑛𝑔, 𝑜𝑟 𝑑𝑖𝑑 𝐼 𝑝𝑙𝑎𝑦 𝑡ℎ𝑒𝑚?)
They confidently said yes. Apparently, this isn’t just an age problem.

That exchange came back to me when I read a 𝐄𝐮𝐫𝐨𝐛𝐚𝐫𝐨𝐦𝐞𝐭𝐞𝐫 finding: only 48% of EU employees said they could recognise an AI-generated fake video. After our little experiment, I wonder how much any of us should trust our own answer to that question.

The confidence trick has been around for centuries. It worked face to face, over the telephone and through email. Now it can arrive on a video call with a familiar face and voice. The tools have evolved; the idea of borrowing someone’s credibility to influence our decisions is very familiar.

𝐖𝐞 𝐤𝐞𝐞𝐩 𝐚𝐬𝐤𝐢𝐧𝐠 𝐩𝐞𝐨𝐩𝐥𝐞 𝐭𝐨 𝐛𝐞 𝐦𝐨𝐫𝐞 𝐯𝐢𝐠𝐢𝐥𝐚𝐧𝐭. 𝐁𝐮𝐭 𝐢𝐬 𝐭𝐡𝐚𝐭 𝐞𝐧𝐨𝐮𝐠𝐡 𝐰𝐡𝐞𝐧 𝐭𝐡𝐞 𝐭𝐡𝐢𝐧𝐠𝐬 𝐰𝐞 𝐫𝐞𝐥𝐲 𝐨𝐧 𝐭𝐨 𝐣𝐮𝐝𝐠𝐞 𝐚𝐮𝐭𝐡𝐞𝐧𝐭𝐢𝐜𝐢𝐭𝐲 𝐚𝐫𝐞 𝐛𝐞𝐜𝐨𝐦𝐢𝐧𝐠 𝐞𝐚𝐬𝐢𝐞𝐫 𝐭𝐨 𝐦𝐚𝐧𝐮𝐟𝐚𝐜𝐭𝐮𝐫𝐞?

This makes me think about how we build organizations. Security needs to run through our platforms, processes, the way we grant authority and our willingness to question a request. Even the relationship between digital and physical deserves a place in that reflection—along with the friction we might accept when the stakes are high.

My kids mistaking a song for AI made for a good family moment. At work, misplaced confidence can have different consequences.

𝑨𝒍𝒐𝒏𝒈𝒔𝒊𝒅𝒆 𝒉𝒆𝒍𝒑𝒊𝒏𝒈 𝒑𝒆𝒐𝒑𝒍𝒆 𝒓𝒆𝒄𝒐𝒈𝒏𝒊𝒔𝒆 𝒅𝒆𝒄𝒆𝒑𝒕𝒊𝒐𝒏, 𝒂𝒓𝒆 𝒘𝒆 𝒃𝒖𝒊𝒍𝒅𝒊𝒏𝒈 𝒐𝒓𝒈𝒂𝒏𝒊𝒛𝒂𝒕𝒊𝒐𝒏𝒔 𝒕𝒉𝒂𝒕 𝒓𝒆𝒎𝒂𝒊𝒏 𝒔𝒂𝒇𝒆 𝒘𝒉𝒆𝒏 𝒔𝒐𝒎𝒆𝒐𝒏𝒆 𝒊𝒔 𝒄𝒐𝒏𝒗𝒊𝒏𝒄𝒆𝒅?

#Cybersecurity #SocialEngineering #Deepfakes