21923 - SCTE Broadband Aug2023 COMPLETE v1 (1)

20 Vol. 45 No. 3 - September 2023 Issue The ’paperclip apocalypse’ theory has been doing the rounds online a lot since ChatGPT was launched in November last year. A grim scenario; AI takes its instructions to create paperclips so seriously it depletes the world’s resources to the extent that the planet faces extinction, at which point its capabilities extend to overriding man’s own instructions. Cut to: humans lose control of the monster they have created. Altman’s concerns are not too far off this; the threat is real. In the wrong hands, specifically outlined during this meeting with the Senate, elections could be rigged, democracies dismantled, misinformation via deep fake AI will be so rife it will be impossible to know what is real and what isn’t. Bad actors and corrupt regimes the world over could potentially influence, direct and instigate all manner of irreversible crimes. It makes Cambridge Analytica’s alleged antics meddling with Facebook a few years ago look like child’s play, especially with a US election coming up in the next 15 months. An example of a deep fake image is below. This is an entirely fabricated image made by Midjourney, a creative AI app, and if you look closely, you will see that Donald Trump has three legs, among other errors, but that has not stopped this image from going viral. The release of this image, at the time of Trump’s indictment in March this year caused what Instagrammers call ‘Internet meltdown’ and a lot of confusion amongst observers of US politics. Eliot Higgins, the creator seemed surprised by the response. “The Trump arrest image was really just casually showing both how good and bad Midjourney was at rendering real scenes, like Trump with three legs and a police belt,” he explained to The Associated Press, who were trying to verify the image. It may well be one of the factors that compelled Sam Altman to conclude, “We think that regulatory intervention by governments will be critical to mitigating the risks of increasingly powerful models,” Altman said. Passing legislation to protect us from all the above moves at a glacial pace however, and by the time it is passed it will already be dated and possibly unfit for purpose. Welsh is concerned. “The Government can’t keep up with this. They’re already so far behind, and they’re not experts. They’re not the right people to be making those decisions. And what they’ll do at some point is introduce something that’s either too heavy- handed or has no teeth whatsoever.” He went on, “I agree with people who say that this isn’t just like the steam engine or internal combustion or computers, even the microchip or radio waves, which have been used for good and evil in all cases, but they’re blunt instruments. AI is not. It’s a very subtle instrument. And again, it could be used for good or evil, or worse, it could be dangerous without intending to be. And that’s where it’s a challenge. It’s a real balance, using it and making good use of it versus being scte long read FAKE Credit: Eliot Higgins