Chat GPT Ate My Hamster

The menace of machines

The internet has been febrile for the last few months with stories about artificial intelligence (AI).

Of course AI has been around for a while, but it is Open AI’s “Chat GPT” that seemed to act as a trigger for the fever, perhaps because its facility of use, and the spectacular speed at which it can spew sentences, which most of the time are of remarkably good quality, the odd “hallucination” or out-of-date information aside.

The outbreak of Chat GPT fever seems to have acted as a catalyst for an arms-race between various companies working on similar programmes, such as Google with “Bard”.

As with the space-race between America and the Soviet Union in the 1960s, to be the first to land a person on the moon, the competition seems to have acted to spur development at an accelerating pace.

Pandora’s Box

The capability of the various AI progammes has been so scary that even some of the greatest movers and shakers in the industry and elsewhere signed a letter a couple of weeks ago, calling for a review and pause, to ensure that humanity can control and understand what it is unleashing-

However like Pandora’s Box, once the lid has been opened, there is no putting AI back in again. At least we still have hope. Or do we? It is difficult for non-experts to know what the risk is, possibly even for the experts.

Predictive text

To the layperson, some of the froth and fever seems excessive. For example all the programme is doing is a bit like predictive text, which has been around on our mobile phones for a while, sometimes with amusing effect when it gets it wrong.

So with Chat GPT and Bard, you start the sentence and the programme uses all its accumulated knowledge from being force-fed the internet, to choose the next word, over and over again, at high speed. For example you start the sentence and it makes up the rest, eg. “Boris Johnson is a ….”

It just fills in the blanks, politely, and with political correctness in the case of Chat GPT. (In case you are interested it completed this sentence for me with “…..a British politician and the current UK Prime Minister”, which was right on the first part, wrong on the the rest), and probably quite different to the answer you might get if you were to ask this question in your local bar.

Losing Luddites

From a non-technical person’s viewpoint my reaction, after initial astonishment at the capability of the programme, is so what?

Well it does matter, as the technology is likely to destroy countless jobs. But in a way that is what technology has always done, ever since the invention of the loom, and probably even before that, back to the first printing press putting scribes in monasteries working with quill and ink out of business.

The Luddites lost their battle against the loom and it would seem that any attempt to resist the worldwide deployment of AI is similarly doomed to failure.

The dangers of AI

The technology it itself is probably fairly neutral — as usual in human affairs it will be used for good, for example to speed up research into diseases and smooth positive interactions, and for evil, for example to invent new weapons or as described in this article, neurotoxins which hadn’t been known of before -

As climate scientists sometimes say, if you are not scared, you haven’t understood the science.

Just because a technology is theoretically neutral, doesn’t mean it is safe, or that spreading it round ever more, will make the world a better place.

Motorway madness

Some of the deployment of AI seems a little, well — stupid, to put it crudely. For example this week the UK government has allowed the Ford motor company to activate the facility for hand-free driving on motorways in some of its vehicles for the first time.

The driver still has to have their eyes on the road all the time, or the facility shuts down, and puts the brakes on. All very clever, but this is the same UK which already has a problem with its “Smart” motorways, which use the hard shoulder as an extra lane. The “Smart” motorway system is meant to make use of cameras which close the hard shoulder if there is a broken down vehicle, but guess what, about ten per cent of the cameras don’t work at any time (this is the UK), which has contributed to fatalities where drivers have ploughed into the back of the stationary vehicle. Add cars with a driver half-watching the road, and it is not hard to see what is coming-

Sinister stalking

There is an interesting and disturbing article published on Medium today, describing how AI can be used to stalk people by trawling the internet for their social media and other information:

Bot bother

Also today there is a report that Twitter is being overrun by Russian bots because Elon Musk has cut costs by sacking the few staff who were moderating content and usage-

Back to the hamster

Closer to home and a subject closer to our hearts, the use of AI generated text threatens the livelihoods and hobbies of millions of writers. There are already countless articles, and even some books, written by AI.

Did Chat GPT really eat my hamster? No of course not, but it is after my Medium earnings, which is even worse.

To join a new Facebook group “Medium Matters” where writers can support each other and share articles.

Mastadon- you can find me here