How I learned to stop worrying and love AI

Prostock Studio via Shutterstock

The movie buffs among you might remember the tagline to 2016’s Independence Day: Resurgence as “We had 20 years to prepare – so did they”.

That’s a bit like how I feel about artificial intelligence.

At least once a week for the past few years I’ve felt the dread of AI – sinister, threatening, malevolent – looming above me and my profession, broadly defined as journalism and “content creation”.

The promise of AI is that it will free us up from the dull drudgery of everyday life – think laundry and ironing shirts – and allow us to concentrate on the creative, life-enhancing side of things, like painting or even writing a column for AGBI.

---

---

The threat is that AI will end up writing the column and force me to the ironing board instead.

But the bullet had to be bitten. I had tinkered with ChatGPT when it first appeared – with disappointing results. In response to my unsophisticated prompts, Chat usually produced a stream of simplistic garbage that could have been written by a 12-year-old with Wikipedia.

What’s all the fuss about, I thought. This will never replace the razor-sharp insight of my own “content”, honed by years of interaction with real humans. 

Nonetheless, AI was not going away. So I decided to enrol at a half-day workshop at the Capital Club in DIFC. Three hours should do it, I thought.

The event was run by Alaa Dalghan, MD of a boutique tech advisory firm called Cognit DX, who turned out to be a master of the PowerPoint presentation leavened with some pretty astute observations about the impending contest between humans and machines for control of the planet.

I felt inadequate and ignorant virtually from the off. “If you don’t know how to do Word, PowerPoint and Excel, you’re illiterate,” he said, looking straight at me. How did he know?

“AI is coming for white-collar jobs – unless you know how to use it,” was the theme of his presentation, reinforced by images from the Terminator movies and apocalyptic messages from experts. “We could witness the first non-organic life form on Earth in 4 billion years”, according to author Yuval Noah Harari.

Also disconcerting was Dalghan’s frequent use of the word “yet”.

“They cannot replicate the human brain, yet” and “artificial general intelligence [AI’s most powerful form] is a theoretical state that doesn’t exist, yet”. I’m sure Dalghan meant to reassure, but the impression I got was that humankind was rapidly running out of time, as in Independence Day.

When he started talking specific details of how to use AI, I found myself on a steep learning curve.

For one thing, I learned that we are not just at the mercy of ChatGPT. There are a number of AI outfits out there, all competing with each other to develop the best software and fighting each other for the graphic processing units (GPUs) that are essential to “teach” AI software in the selection, curation and categorisation of data.

OpenAI/Microsoft, Meta, Google and now Elon Musk’s X are in a race to perfect and bring to market their own AI technologies.

One of the most intriguing, as Dalghan told it, is Claude, the brainchild of two ChatGPT dissidents who left the company in protest at the hook-up with Microsoft. Claude is winning plaudits as the best AI currently in operation, according to experts.

The other thing I learned was the essential need for good prompting. AI responses will only be as good as the questions you put to it. Dalghan imparted several useful tips to optimise prompt efficiency, for example by using the word “please” at all times.

“Always start your prompts with ‘please’. If they end up ruling the world, we should be polite to them,” he explained with a wry smile.

I did not stay until the end of the presentation, which moved towards visual and graphic forms of AI, mainly because I don’t think I’ll really need to create pictures of brown French poodles driving Teslas in my everyday work.

When I left, I genuinely felt a bit less frightened by the prospect of AI. You fear what you do not know, and I had made a good start on taming and harnessing AI to work for me.

There is no chance AI will replace humans. Yet. Is there?

Frank Kane is Editor-at-Large of AGBI and an award-winning business journalist. He acts as a consultant to the Ministry of Energy of Saudi Arabia and is a media adviser to First Abu Dhabi Bank of the UAE