Before I start ranting about the latest AI shenanigans, let’s all take a deep yoga-like breath in, and out… “phoooooooooo”. I’m really genuinely happy that we had at least one month of peace and quiet in the bumpy AI world.
I’ll start off with this: AI will NOT erase the human race in 2030, 2090, 5080, or any other year. In fact, this is getting really boring and dull. And if you notice, it’s always the same companies, and personnel from the same companies that keep going on with this theatrical dramatic show. Every single time!
You don’t hear this from the older and more mature tech companies. Actually it’s only the 2 cousins – OpenAI and Anthropic. They’re the real “drama queens”. I mean imagine a world without those 2 companies, how boring it will be, aye.
So we’re beginning to lose focus here, and we need to remember that everything that happens, happens with a motive and an underlying reason.
I still can’t wrap my head around the fact that a company that makes its living out of building frontier AI and is hell-bent towards AGI keeps on banging on about AI erasing the human race, or taking over people’s jobs, or some other great catastrophe coming from AI. If they really believe all this, why are they still in the AI-making business?
Anthropic in particular is just a company I fail to understand. They sincerely believe that AI’s capabilities are really scary. This came across to me again when I was in Washington DC at a working session with US regulators and physicians organised by the American Medical Association a couple of months ago. The panel I spoke on was shared with Google, OpenAI, and Anthropic later on (and a few other companies). And the guy from Anthropic soberly stated that “we don’t know what we’re building, we don’t know the implications of what it could do. We don’t know what we’re dealing with here”.
And I’m sitting there, thinking to myself, “then why are you guys still racing to win the AI race, and keep on building frontier AI models. Why are you developing what you don’t know or understand?” Or I’m I missing something here?
And now we have yet another call to pause AI development, and slow down AI development from Dario Amodei, Anthropic’s CEO; and for the first time, his old buddy Sam Altman agrees with him. (Maybe they’ll be holding hands at the next AI safety summit!)
This is yet another dramatic showdown from American AI start-ups. All in the race towards IPO.
We mustn’t forget that since 2023, there have been several statements, pledges and promises to slow down the pace of AI development. And it’s slowing down alright. So much so that so-called rogue agents keep on hacking into people’s websites, and even hijacking them! The latest one was from Open AI where an agent hijacked a German website and turned it into a bulletin board for other agents. Of course Open AI kept this under wraps.
You know when I saw these latest headlines on the AI Apocalypse, I first tried to imagine how AI will actually erase us. Does it have hands and feet to carry out the work? Oh, but I’m being a bit obtuse here aren’t I? Because it’s the humans developing and launching these AI agents and models that will destroy the human race if left unchecked. Not AI, don’t blame it on AI.
So this leads to the next question on preventing these sort of wild things from happening. Like thinking along the lines of cybersecurity, AI safety and alignment, Responsible AI, AI governance, AI regulations and so on.
Before I go any further though, it’s important to remember that these folks are all from the United States. So if they decide to slow down AI development, China is going to win the AI race. Because I’d like to see how Trump will whisper sweet nothings to Xi Jinping’s ear, persuading him to stop all things AI. That’d be the joke of the century.
AI has a lot of challenges that have remained unresolved over the years. The only “progressive” advancements we’ve seen are new AI models, new competing AI agents and of course new AI hype and scares. Like the latest model from OpenAI, GPT 6 that has now been declared to be AGI (of course it is).
We’re still dealing with the never-ending issues around AI’s violation of human rights, so much so that the UK’s MP and Lords have called for a new law to address AI and human rights. Need I remind you that while AI is still quite good for day-to-day management and tasks, it’s still a terribly-built technology.
AI is driving further inequality across the world, with a concentrated oligopoly and power control. It’s still full of life-altering biases, representational harms and provides misleading results and hallucinations that have led many people to their death.
How about the compounding energy crisis? People in London, particularly in Slough are suffering from increasing unliveable temperatures due to the presence of data centres in that region with temperatures raised by up to 9°C. This is just one example among hundreds across the world.
What do you think caused the flash floods in Nepal? Or the ongoing wild fires across many parts of the world, destroying homes and claiming lives. You guessed right a key factor is the exacerbating climate crisis, highly driven by power-hungry data centres.
There are also amplified biases in hiring, credit, and justice. We’re facing risks of cultural erasure, data colonialism and control to underpaid annotators. What about human surveillance, AI in military use being used to destroy civilians and innocent children, etc.
And the latest, which has always been there actually, is AI misdiagnosis and hallucinations in healthcare. While the AI note-taking industry has flourished and been adopted across several industries including healthcare in the form of Ambient AI, people are now getting misdiagnosed, receiving wrong medications that were swapped for a different one with a similar name, repeat prescriptions removed from summaries, hallucinations in notes where drugs that were never discussed were prescribed etc.
This is scary stuff. More scary than unknown, unseen and unproven AI threats. So if we’re not careful, while everyone is listening to the incredible actors who are probably in the wrong industry; inaccurate, untested and unsafe AI technologies will be silently erasing human lives through healthcare.
Don’t get blindsided folks. Let’s focus on what really matters. It’s high time regulators from across the world step up to ensure companies are held accountable and have the right AI safety measures in place. Tackling these issues by one risk / harm at a time is the best approach.
For the rest of us, rest easy – you’re not going away anytime soon.
