Mark Zuckerberg, chief executive officer of Meta Platforms Inc., during the Acquired LIVE event at the Chase Center in San Francisco, California, US, on Sept. 10, 2024. (David Paul Morris/Bloomberg via Getty Images)
Artificial intelligence is developing at a rate too fast for legislators to keep track of and protect citizens, including children and teenagers who are being negatively affected by the growth of fake content and falsehoods peddled on social media.
Social media and the use of AI to create fake online images have become “the tobacco of today”, said Ajit Gopalakrishnan, an AI thought leader and head of Odin Education, a division of Jendamark Automation, a global tech company that provides solutions to the automotive sector.
At a recent workshop on AI in Durban, Gopalakrishnan gave the example of a Vodacom advert for a Samsung Galaxy phone, which uses a dull image of a lifeless-looking young woman holding up a phone that shows a photo of her posted online that exudes colour, confidence and joy.
“We are currently living in probably the highest-ever rates of teen depression, and look at advertising.
“In real life [in the advert], look at her expression. She’s dull. She’s boring. She’s black and white and online — how does she look in the same moment?”
The advert for the phone harks back to the tobacco advertisements of the Eighties and Nineties, which portrayed images of successful people smoking, happy and enjoying the leisure pursuits of the rich such as high-end parties and sailing.
Gopalakrishnan highlighted eight risks of AI to society, businesses and humanity, telling the workshop: “Essentially, if AI didn’t develop at all from today, these risks are really prevalent. This is not about a future risk. This is a current risk.”
Among these are the threat of polarisation, bias, joblessness, loss of intellectual property, dispersion, mass manipulation, deepfakes and the inability of legislation to keep pace with AI.
Social media and AI are already being used to polarise society. Extreme content on the political left and right drives traffic and keeps people engaged for longer, while content expressing moderate views is not popular as it does not induce emotion and is therefore not pushed in posts, Gopalakrishnan said.
AI also carries the inherent bias of its creators.
It has the potential to replace many workers across all types of jobs from accountants, scriptwriters and Hollywood actors to electricians, as robots could replace them.
“It has been shown that the jobs of plumbers are probably the safest because machines can’t yet perform their skills,” he said.
Gopalakrishnan said CVs are already being reviewed by algorithms to exclude those without certain keywords.
This is why the European Union AI Act has attempted to combat this by imposing a wide range of obligations on the various actors in the lifecycle of a high-risk AI system.
“For example, AI should not be allowed to make decisions because humans cannot inspect the decision-making process.
“People have a right to an explanation and therefore any decision that has a significant legal effect cannot be delegated to an algorithm.”
Gopalakrishnan said this has been touted as the last year of the human election as AI can be harnessed to survey and assess the psychological state of the electorate to inform the promises made by political candidates during campaigns, while deepfake videos and voice impersonations pose new threats of fraud.
“Imagine, you know, if I were to call my mom with my voice and my face (showing on the mobile phone) and ask for her pin number and her bank details; she would do it.
“This is the reality we are in today. I think people are just clueless as to what is coming our way,” he said.
There were also concerns about the use of AI in weapons development and warfare.
“Russia and America, the two countries that cannot agree on anything, agree on the fact that everyone must have lethal, autonomous weapons so, in all the wars right now, in Ukraine, Middle East, Israel, Palestine, AI is making kill decisions.
“In fact, one of the things that we know is the Israeli government actually uses AI to determine targets using databases,” he said.
Gopalakrishnan believes that the time will come when people no longer believe information published on the internet because of the proliferation of fake, AI-generated content.
Estée Cockroft, founder of Screen Smarts, which coaches people on how to be street smart on screens said it is “the Wild West” when it comes to cyberspace and legislation.
“Not only is legislation far behind ring-fencing where we are currently standing in terms of invention but technology is accelerating away at pace,” she said.
“It is moving too quickly for policy to catch up.
Cockroft suggests that, because of the negative effects, children and young teens should not be given access to social media.
“We have seen a dramatic rise in social-media depression globally. Experts are referring to it as a ‘global adolescent mental health crisis’.
“Not only does comparison give rise to unattainable ideals but children and adolescents desperately need what technology cannot provide — close connections with their family and friends,” she said.
However, while there are more voices taking companies like Meta to task regarding fake posts, accountability and control remains “a long way off” and “there is zero transparency on their parts”.
“What we have shared in the past can be manipulated today with technology that wasn’t even invented when we posted the content. And this with very little recourse,” Cockroft said.
She regularly hosts workshops on AI at schools, focused on topics such as polarisation, deepfake technology, source interrogation, the danger of AI nudes and protecting your privacy online.
“Open conversations on these subjects are urgently needed …Generally, Generation Alpha is more sceptical than previous generations.
“However, it is imperative to include critical-thinking skills in all digital citizenship education.”
Cockroft also foresees a time when generative AI will control most media channels and people will no longer believe what they read and see online.
“As humans, we will be sceptical of all content because we will be unable to discern real from fake,” she said.
She concedes that technology has a place in a child’s upbringing.
“It is necessary for our children to use AI and learn robotics. It is vital for them to experiment with coding and learn through discovery.
“However, they do not have to have smartphones or social media to be tech-savvy or connected.
“Smartphones should be delayed until high school and no child should be on social media until they are 16 years old.”