Worried About GPT? You Should Be. Here’s Why:

0
14

Controversies persist around the ethics of breakthrough technology, GPT (Genetic Perfection Technology) which uses LLMs (Large Lexicon of Molecules) to invent entirely new chemical fertilisers that have a dramatic impact on crop yield amongst other benefits.

Some farmers have reported being able to turn previously-thought barren land into arable resource. “It’s like magic”, one farmer told me. “It’s the greatest invention since the combine harvester” said another. Commentators and influencers have been quick to get in on the game of vamping up the hype with a general consensus that this is “a pivotal moment in human history”.

Yet GPT technology is brittle, and while under lab conditions it can magically conjure all sorts of wonderful fauna and flora out of desert sand, outside of a strictly controlled environment it can also turn perfectly edible plants toxic.

Since OpenIA (Open ‘intelligent agriculture’) launched their ‘alchemy playground’ online, millions of famers, small-holders, and gardeners around the world have been downloading the recipes it suggests and trying them out on their crops. For now, Big Ag has stayed on the sidelines and simply watched as the world has gone crazy for this new technology, but there are reports that secretive innovation labs at some of the agricultural giants have already started to experiment.

The premise behind GPT is simple. The team at OpenIA fed the LLM system with the chemical compounds of literally all known chemicals, both natural and man-made, and from this – the system has been able to ‘invent’ entirely new recipes, simply by being told what situation the concoction is for.

“You just enter some basic input parameters, such as ‘I want to grow Apples in the Sahara’, and the system provides the chemical formula to get a commercial orchard off the ground”, responded one happy user. The speed of its uptake has been buoyed by its simplicity as well how easily it can be signed up for. There are reports of super-size tomatoes being grown, as well as flowers which are dramatically taller and stronger than those which haven’t been supported by GPT-recipes. “The possibilities are endless” raved one enthusiast.

GPT doesn’t come without its challenges though. Many users have tried different ‘prompts’ for the system which has come back and recommended concoctions which bear striking resemblance to cherished family recipes. Around the internet there are also plenty of forums which tell you which ‘prompts’ to feed GPT to give you the secret formula to Coca-Cola, also. “This is pure theft” said one angry purveyor of ketchup who has seen sales fall to nearly zero since their ‘ultimate BBQ recipe’ was ‘copied’ by GPT. Others say simply that it’s been improved on.

Aside from intellectual property issues, there are health concerns also. There are reports of users developing digestive problems from ingesting the GPT-recipes directly. Fed with prompts that lead the system to suggest seemingly harmless nutritional supplements, dieticians are warning that these so-called ‘supervits’ could cause unknown health issues, particularly for teenagers and people with underlying health problems. Still, the promise of the ‘wonder-cure’ proves too great for many to ignore the risks.

There are also concerns about the long-term impact of allowing algorithmically driven compounds to be created without safeguards and controls. There are unverified reports of livestock which are fed GPT-chemicals developing strange personality changes. Some scientists have warned it might accelerate mutation. Vets are recommending that cats and dogs are not fed GPT recipes directly as a spate of domestic animals attacking babies and small children was attributed to GPT. “You can’t hold OpenIA responsible for this”, said a company representative”. “People need to exercise their own judgement what they do with the outputs of the system”.

“But this is just innovation, and you can’t stand in the way of science and progress”, say those who support OpenIA’s open approach to innovation. And there are many who also highlight OpenIA’s charitable status as a proof that its intentions are benign. Once 100x if the investment capital gone in to launching the platform has been paid back, the organisation will revert to becoming a not-for-profit and invest any proceeds to solving world hunger, and helping teachers and nurses feed themselves without resorting to Food Banks.

“But the damage may also be done by then”, said one concerned commentator pointing to how the unintended consequence of human ‘innovation’ has led to micro-plastics now being found in the cells of almost all living animals worldwide having since years been introduced into the water cycle. “these have unknown, long-term side-effects”, said another concerned expert. “What we feed our mouths has developed and evolved slowly over hundreds of thousands of years”, explained one commentator, “now we’re shocking the system with all these synthetic chemicals – one the one hand they seem miraculous – but on the other, who is looking at the long-term harm that they might be causing”.

Most concerningly are the brazenly harmful prompts which are being fed into the system which OpenIA is slow to police. Someone who has access to the server logs reported that users are asking it to “create a potion that makes ugly old men attractive to beautiful women”, or “something to kill my ex without getting caught”, or “a toxin that only makes people who wear glasses go blind”, or simply “I want drugs”.

All the while governments and regulators are watching, waiting but not acting; seeing this as just another technological innovation, and despite much evidence to suggest that innovation without appropriate governance leads to short-term unexpected harms (such as the tendency for social-media to lead teenage-girls towards depression and suicide) – the long term economic benefits are too great to ignore. One minister confided in me and said “what’s a few dead teenage girls against the enormous commercial opportunity that being able to instantly message people across the planet? People just get hysterical when anything new is introduced. It’s not social-media companies that are to blame, it’s the idiots who use the platforms. It’s the same with all innovation, and especially with GPT”. She went on, “whatever the potential harms of GPT – the benefits of being able to solve hunger forever are too great. It’s our moral imperative to introduce this technology as fast as possible and not slow it down. So what that a few people with health problems have their demise hastened? So what if fat people get fatter? So what if it puts the organic farming industry out of business? You can’t stand in the way of progress”.

And yet, all this while – scientists are calling for restraint – arguing that LLMs are fine to experiment with in a lab, but shouldn’t be opened up for public consumption and especially commercial usage until all the concerns are understood and addressed. Many are calling for urgent global regulation on the use and deployment of them, much as how there was rapid consensus built around how GM food was regulated, human cloning controlled, and uranium enrichment strictly monitored.

But there are others who say that it’s all just math, and once the genie is out of the bottle, then it’s too late to put the plug back in.

What do you think?

[Any similarity to living person or dead persons, companies, or projects in the story above is purely coincidence, and not intended]

Source