Editorial: AI uncertainty as ChatGPT has a birthday

The OpenAI logo is seen on a mobile phone in front of a computer screen which displays output from ChatGPT. Photo / AP

Editorial

EDITORIAL:

The first publicly-available artificial intelligence celebrates its first birthday at the end of this month: ChatGPT (Chat Generative Pre-trained Transformer), a large language model-based chatbot developed by OpenAI, launched on November 30, 2022.

In the months that followed, there were countless reports of how the new technology could transform our society for both good and bad.

By removing the human element, systems could work with a higher degree of accuracy 24 hours a day, seven days a week, reducing costs and improving processes. The potential in areas such as medicine and healthcare, farming, transport and even in tackling existential problems like climate change are exciting. It’s even been suggested that AI could referee a Rugby World Cup.

But there were claims that AI could steal jobs. A report from Accenture suggested around 40 per cent of jobs could disappear, mostly in clerical and secretarial roles. The upside would be the growth in machine learning specialists, data analysts and scientist, and digital transformation specialists.

A strike in Hollywood earlier this year was felt around the world, including in New Zealand, amid fears that studios would replace writers and actors with AI-generated words and pictures. Several big-name stars added their voices to the cause, such as Scarlett Johansson, Tom Hanks and Keira Knightley, who looked to copyright her own face.

A number of big-name tech entities raised the spectre of the “Terminator scenario” — a future where machines take over and threaten human existence. The Centre for AI Safety in May warned of the weaponisation of such technology, the spread of misinformation, the concentration of power into fewer hands and the over-reliance on computer systems. Other specialists have countered that the existential threats are overblown, distracting from the real near-time problems of AI.

“Advancements in AI will magnify the scale of automated decision-making that is biased, discriminatory, exclusionary or otherwise unfair while also being inscrutable and incontestable,” Elizabeth Renieris at the Oxford Institute for Ethics in AI told BBC News. The increase in misinformation, erosion of public trust, the entrenchment of social division and the “free ride” on collective human endeavour are the problems today, she added.

The calls for limiting the use of AI have prompted governments into action with the White House publishing its Blueprint for AI Rights.

The British Government last week hosted an international summit resulting in the Bletchley Declaration signed by 28 countries, including the US, China, the EU, India and Australia, agreeing to work together on AI safety research.

Elon Musk, owner of X (formerly Twitter) and Tesla, said from the summit sidelines: “For the first time, we have a situation where there’s something that is going to be far smarter than the smartest human ... it’s not clear to me we can actually control such a thing.”

We have been warned.