Nicky Marr: Should we worry about robots?
Register for free to read more of the latest local news. It's easy and will only take a moment.
Ask me what I think of artificial intelligence and sentient robots, and I glaze over. I grew up without computers, and the success of my day depends on whether I can get my printer to work.
Sentient robots and chatbots are still the stuff of science fiction, aren’t they? R2-D2, The Doctor’s K-9 companion, and Holly from Red Dwarf.
Or are they? We have a driverless bus between Inverness Campus and the retail park, and Japan has been using automated humanoids to care for the elderly for decades. Robots are commonplace in manufacturing, doing menial tasks without the need for breaks, training or sick days. These aren’t quite sentient – yet – but they are replacing humans in jobs, and that is inexorably changing our industrial landscape. In a society where we want cheap goods and meaningful jobs, that could be thought a good thing. But how far will the balance tip? And what happens when the robots and chatbots do become sentient, and can think for themselves? And can learn, solve problems and make decisions they’ve not needed to be programmed for? A flurry of headlines last week about the dangers of AI meant I couldn’t dig my head in the sand any longer.
On Monday, the man whose research led the way for bots such as ChatGPT, quit his post at Google so he could warn the world about the dangers of Artificial Intelligence.
On Thursday, warnings came both from the UK Competition and Markets Authority and from The White House, that the AI sector must be reviewed before the intelligence of machines outstrips the intelligence of the humans who created them.
And an open letter asking for a pause in AI development was signed by 1000 industry leaders including Elon Musk, whose company owns ChatGPT, Steve Wozniak, the founder of Apple, and engineers from Amazon, Google, Meta and Microsoft.
That all adds up to something we should start paying attention to. So what problems could runaway AI cause? The spread of misinformation is one. Job losses is another. I decided to test ChatGPT on both.
To test the spread of misinformation, I typed in “Who is Nicky Marr?” Narrowing the search to the Highlands, it found me.
Some of its information was spot on. It knows I’m a broadcaster and journalist and flattered me with the description, “a prominent figure in Scottish media and culture”. (Aye, right.) But otherwise it was rubbish. It credited me with having written a string of business and political books I’ve never heard of and which, as far as I can tell, simply don’t exist.
Lots of artificial, then, less intelligence.
Click here to read more from Nicky Marr
But could AI replace my job as a columnist? I asked it to write this week’s column for me. I gave it my word count and clicked “Generate Response”. Reader, it refused.
So I asked it instead for inspiration. “What should I write this week’s column about?” Surprisingly, it didn’t generate the title ‘Yikes! Could ChatGPT replace me as a columnist?’, but instead suggested I write about the impact of climate change, the independence debate, or the challenges facing small businesses. Interesting. Has it been reading recent editions of this paper?
Still in trial stages, it learns from every question, and from feedback to every answer it generates.
But can it create beautiful, stirring poetry? Can it be creative? I fed in the title “Inverness Castle” and asked for a poem.
Poets may be safe – for now. My P4 teacher would have returned this with “see me” scrawled in red pen:
Inverness Castle, mighty and old,
Stands atop a hill, strong and bold,
A symbol of Scotland’s history and pride,
A sight to behold, far and wide.
