My children are not AI – #5 They both don’t know what is real.

My children are not AI – #5 They both don’t know what is real.

Tuesday, December 19, 2023 ChildrenAI

Christmas is looming, and Santa Claus is getting ready for his gifts delivery… at least that’s what my son believes! Knowing what is true or fals is not easy and AI does not help! In this 5th article, I explore how my son and AI manage (or not) to tell facts from fiction.

My 5-years old son is a big fan of Harry Potter. We spend hours reading books, discussing characters and casting spells. One day, he wanted to verify  that magic wasn’t real. I confirmed, but he immediately added “except for Santa and elves!”. To be honest, it made me a bit uncomfortable. I try to instill in my son a sense of honesty and critical thinking, but I fully participate in the biggest plot of all time: that of a big old man flying in the air and offering presents to good children all over the world, and all in one night. If this is true, then why can’t Harry Potter be real too?

I asked ChatGPT if Santa is real and it replied: “Santa may not exist in a literal sense, the joy and happiness that Santa represents certainly do exist in the hearts of many during the holiday season”. Does it mean AI knows better than my son what is true or false?

We have seen in the news several stories about people getting responses from ChatGPT that were wrong, like this lawyer in New York citing 6 bogus cases and now facing potential sanctions for unethical practices. Generative AI was not made to tell the truth, but to give probable answers based on the data it has been trained on. On the internet, there is more than enough content about Santa to enable ChatGPT give me a sensible and quite poetic answer. However, if you ask something far more complex and specific, it will make up a plausible answer, but not necessarily a true one.

We now live in a highly connected society with social media playing a big role in keeping us informed about what’s going on in the world. For the younger generation, it plays an even bigger role than more traditional media. With this significant change in our communication practices, we have seen an unprecedent increase of fake news these last few years. Research from Carnegie Mellon University indicates that “45% of tweets about coronavirus are from bots spreading fake information […] aiming at sowing division within America”. Fake news spreads 6 times faster than true news, because their sensationalism makes them more appealing and shocking, leading them to be shared and reposted more often. I strongly encourage you to consult other interesting stats and facts about fake new, as well as on other impacts of technology on humanity in the Ledger of Harms published by the Center for Humane Technology.

Fake new is becoming more and more plausible with the recent development of deepfake photos and videos, enabled by AI. Many examples can be found, from Pope Francis wearing a white puffer coat, to Mark Zuckerberg talking to CBS News about the "truth of Facebook and who really owns the future", or celebrities’ fake porn. They are being used for propaganda and in the disinformation war in armed conflicts like the Russia-Ukraine and the Israel-Hamas wars. AI can even animate pictures of dead or fictional characters like Salvador Dali or Mona Lisa. It is getting harder, and sometimes impossible for humans to tell whether a photo or a video is even real.

On the other hand, AI can also help identify deepfakes much better that humans can. Adobe developed new feature in Photoshop to change facial expression on photos and on videos with Project Morpheus, but also in reverse to help identify deepfakes and even revert modifications. However, these technologies are not designed to work at scale and cannot keep up with the flow of videos published constantly on the internet. Technology can also help add digital tags to original content and attach metadata about the source of information. Recently, we talked a lot about the jobs that would soon be replaced by AI, journalism being one of them considering generative AI’s impressive writing skills. But journalists and media will have a significant role to play in the near future to verify the source of information and undertake investigations, not just copy or re-write random news they glean on the internet.

So how can I prepare my innocent child, who still believes in Santa, for this world where he will be constantly bombarded with unreliable information? First, he needs to learn not to believe everything he reads, hears, and sees. He will need to control the emotional impact all this news has on him so he can think rationally. He will need to develop a strong critical analysis skill to question the veracity and assess the social context of each piece of news. He will need to find trustworthy sources of information and maybe even need to use new tools for that. And as parents, we have a huge responsibility to lead by example in our everyday life, including telling them the truth… but maybe after Christmas!

No comments yet
Search