 |
| -Image by Gerd Altmann from Pixabay |
This is the post i know some of you have been waiting for. Did you guess which parts of the Body Image post were written by AI? It was the top two bullet point sections – Paragraphs 4 and 5, though I may have automatically tweaked the spelling to Anglo -Australian in places. In the process, I have seen the seeds of my own demise as a scribe. AI just did in moments what it normally takes me a week or two to achieve. My daughter has just written 10 blog posts in less time than it took me to write one and it has even done it without typos and with more logical exposition than mine usually have. Imagine what it will do for report writing and perhaps this is why News Corp is sacking so many of its journalists, though falling revenues may have a lot to do with it. Musicians aren’t immune either. One program can do fake Bach at the drop of hat. I can also see why schools are concerned. Many have banned the use of ChatGPT because it takes all the labour out of doing your own research and putting together a coherent essay.
Meanwhile, the whole family is experimenting with various aspects of it. Oldest son has been using it for technical analyses for quite a while. Youngest son is using it to streamline his strategy in a computer game. At the moment it’s like the early days of the home desktop computer. No one knows exactly how to use it or where it will lead, but the field is wide open and it feels like we are on the edge of another revolution.
The Promise
Imagine not having to hunt for legal precedents in a law firm? Juniors could instead be deployed in the taking of statements or planning strategies. Doctors could let a Chatbot ask the basic questions and point to some likely diagnoses and treatments, while they could instead turn their attention to the emotional disposition of their patients. Architects could come up with environmentally sustainable solutions to housing which involve minimal waste and use of recycled materials. Chefs could plan meals more precisely with a lot less waste. There will be uses which we cannot as yet imagine. Combined with 3D printers they could take the number of homeless people directly from the Census figures, or after a disaster and automatically start printing shelters. There is talk of using AI to replace animal testing in laboratories. The possibilities seem limitless. Here’s AI in use in medical applications. If Neurolink has its way, we’ll be able to upload directly to a computer or a USB and get more accurate diagnosis and treatment.
But …..
The Pitfalls
Did you know we already use AI whenever we engage with a Chatbot on someone’s website? Yes, I know it’s annoying and they are as yet imperfect, but for the operators, they can save a lot of time by dealing with routine matters, FAQs and so forth, leaving only more difficult problems for humans to deal with. Expect these to come nearer to perfection as the technology improves.
Who’s driving the Bus?
It all presupposes that there will be a benign entity behind it and that it will be used for good, but on dark days it’s not hard to see what could happen if the power to manipulate out thoughts, beliefs and ideas could fall into the hands of some dictator. Already ChatGPT has been used to reverse engineer chemicals with negative results and what is to stop people using it to make weapons? I’m glad you asked. Since the chemical fiasco, operators have been quick to insist on a code of ethics for ChatGPT, which I might add has been sorely lacking in some other areas of public life of late – I won’t mention names.
That is not to say that commercial uses and military applications won't overshadow other uses or that criminals won't find new loopholes to exploit which we will then have to scramble to close, just as in online banking or romance. The only reason the Internet works at all is that it was so widely dispersed and started out in the universities as a public good, not as a proprietary product. Nor would it have developed so many uses from knowledge sharing to storm prediction not to mention its many commercial uses, had it not been allowed to evolve largely unmolested in its early stages.
Loss of skills
As to school children being less inclined to develop traditional academic skills – a frequent complaint already with so much being done on keyboards, perhaps it will be more about learning the right questions to ask and how to judge a piece of work and acquired knowledge. Is it cheating to use a calculator instead of adding strings of numbers, provided you know how to do the basic work? Perhaps we need to limit it in the Primary years while children learn the basics or reading, writing and ‘rithmetic, but I think we would be limiting ourselves by banning it completely, especially as it may give way to other forms of knowledge and creativity. Engineers must learn how calculations are arrived at, but they no longer need to work out each one with a slide rule. There is such a wealth of knowledge being produced in every field every day, that we are no longer expected to hold all the details in memory or learn it by rote because we have machines which do that. Perhaps not having to spend time assembling and organising knowledge will free up other areas of the brain for who knows what new skills which we won’t recognise until we get there.
Attribution
At present there is no obvious way of knowing where material has come from, whether it has all been derived from ChatGPT or whether the sources are credible. In the first instance, there should be a marker or imprint indicating that ChatGPT has been used and a list at the end or within of the sources. Nor is there any acknowledgement of the original creator of the work which has been used. It is early days yet and we don’t want to nobble the system when it has barely begun, but particularly in the case of commercial return on reuse, there should also be provision for a royalty system which is consistent with the extent of appropriation.
Loss of Original Voices
If ChatGPT can churn out material at will and even mimic my voice - with similar vocabulary, hesitations, typos and digressions, how will we know when we are dealing with a flesh and blood human and not an algorithm? It can also produce as yet imperfect fake videos which are getting so good that only experts can pick the difference. That’s something to ponder for the future. I wonder if, when AI does most of our written work, will we begin to value those fragments, the poorly constructed essay, the accents and the voices a chatbot hasn’t yet cloned, as truly authentic and original?
Will no one think of the Carriage- Makers?
Did Henry Ford consider what would happen to the buggy makers when he streamlined car production, or the many farriers and saddle makers? I doubt it. Nor would it have stopped the onslaught of the horseless carriage. In the same way, as with the takeover of the office by computers, we may be able to delay the process but not stop it. Did we mourn the loss typists and armies of filing clerks, of the Beta Video machine makers, or the video and music store operators, the Swiss watchmakers or the celluloid photo developers? Possibly, but did it stop the new innovations taking over? Our competitors in the marketplace will use AI to save time and money and to produce a cheaper product. For this reason, we would be far better getting ahead of the curve and incorporating it into our thinking and products.
In turn we should be looking at every single job where AI could be applied and start reskilling those people and planning for the future, including training people in the use of AI. We do not want a great pool of unemployed people as we had after most manufacturing went offshore and so many labour intensive industries became automated.
AI is not going to go away. It is already the subject of a new global arms race and we will struggle to keep up with developments if we do not engage. We have to make sure too that the positive benefits outweigh the negatives and that issues such as loss of privacy and potential for exploitation are considered in legislation before its use becomes widespread, but we ignore AI at our peril.
Another Experiment in AI
The text discusses the potential benefits and drawbacks of AI. On the one hand, AI can automate routine tasks, such as legal research, medical diagnosis, or meal planning, freeing up professionals to focus on more complex or emotional aspects of their work. AI can also contribute to solving societal challenges, such as homelessness or disaster relief, by generating 3D-printed shelters. On the other hand, the text raises concerns about the risks associated with AI, such as the possibility of misuse, abuse, or manipulation of personal data or public opinion by actors with nefarious intentions. The text also questions the impact of AI on human skills and values, such as the ability to judge the credibility and authenticity of sources or to appreciate the uniqueness of human voices and perspectives. Overall, the text acknowledges the potential of AI to transform many aspects of our lives but also calls for ethical and responsible use of AI, as well as for preserving and cultivating human creativity and diversity.
I don’t know about you, but for the moment I prefer mine, with all its faults, but ChatGPT does make an excellent executive summary out of it which would save a lot of time for those who don’t know whether they want to read more or not. Next time I’ll put that up front -duly acknowledged of course!
Comments