______________
WHAT’S UP!
During the recently concluded third edition of the Kampala Writes literature festival, one of the most heated discussions centred on the existence and future of Artificial Intelligence (AI). Titled Can Machines Imagine? Creativity in the Era of Artificial Intelligence, the discussion soon moved beyond what AI meant to creative writers to whether it was a clear danger to humanity.
Opinions were divided, with one of the panellists, Felix Kitaka (who claimed to have built a Luganda AI), insisting that AI is here to stay, and we just have to learn to deal with it. Bernard Nakireza (who runs the Barefoot Law organisation with the help of AI they named Winnie) said AI development is so fast we don’t really know where we are headed.
Especially worrying was some companies using AI to develop more advanced AI, with no human involvement or control. Kenyan author Betty Kilonzo insisted that humanity will persevere, although we have to deal with it right away.
The audience was equally divided, with some declaring AI a useful tool we cannot do without, and that everyone should get on board or be left behind. Others were convinced that AI will sooner or later take over, and that will be the end of humanity.
Since then, almost every day when I check my news feed, someone somewhere is sounding a warning that AI developers are more concerned about competition than with the threat it poses to humanity.
Former employees and whistleblowers of AI development companies are warning that not enough controls are being considered as more powerful AI is developed, which, of course, the companies deny and claim that controls are in fact very much in place.
Nevertheless, a former researcher with AI firm Anthropic claims that AI can kill off all humanity within the decade.
In steps the President of the US, Donald Trump, who dismissed the fears and said any threats from AI will be dealt with by a ‘smart president’. No prizes for guessing whom he meant. So, humanity’s fate is supposed to lie in the hands of the man who thinks climate change is a hoax, and that drinking bleach could cure COVID-19. At this stage, we probably have to decide who’s the greater threat: AI or an increasingly out-of-touch 80-year-old in the White House.
During the debate at Kampala Writes, it was inevitable that someone would bring up the 1984 film, The Terminator, and its sequels. The series is about a fight in the future between machines that have taken over the world and a remaining bastion of human beings, led by a one John Connor.
The machines are led by an AI named Skynet, which sends a cyborg disguised as a human being back in time to kill Connor’s mother before she can give birth to him.
Director James Cameron tells the story of an AI that is given control of almost all manufacturing systems, including military ones, to remove human error and delay. But somewhere along the way, it becomes ‘self-aware’. When operators realise this, they try to shut it down, but it is too late. Skynet sees that attempt to shut it down as an existential threat, and it determines to wipe out all humans. Is that a prophecy, or what?
So, I asked META AI if AI is indeed a threat to humankind. It replied yes, but in a roundabout way. It classified the threat in four parts: capability amplification (where scams and fraud, together with mass surveillance and systematic oppression by governments, are made easier by AI), concentration of power (where companies or countries that control AI systems inevitably control the world, and this is probably the real race between China and the US), loss of human agency (where we let AI do everything for us), and misalignment (where Skynet becomes real).
It concludes that yes, the threat is real, like climate change is real. “... not a movie robot uprising next year, but a powerful general-purpose technology that, if deployed without checks, makes other bad outcomes much more likely, and that we have a window now to shape’.
Note that is used ‘we’, as if it’s human. ChatGPT agrees that the threat is real, but “... ‘AI will destroy humanity’ is far from an established prediction.”
It explains that the real danger is telling a ‘sufficiently capable autonomous system’ to ‘achieve X at all costs’. That is where a Skynet-like scenario will develop. Sam Altman, CEO of OpenAI, insists his company and others like it will do the right thing when it comes to AI development.
“The world should trust that we are going to do the right thing because it’s the right thing and we feel the magnitude of this,” Altman was quoted as saying.
But as almost all of us know, most dystopian stories and movies are based on science gone wrong or on things escaping from labs. There’s still a widely held belief that COVID-19 escaped from a Chinese lab.
So, does anyone really believe what Altman is saying? I don’t, and I think that the chances are great that the Terminator is coming. And hopefully I’ll be around to fight alongside the John Connor of our day.
You can follow Kalungi Kabuye on X: @KalungiKabuye