By Lee Capocchi
A long time ago a scientist and author, Isaac Asimov, suggested it was cheaper and more efficient to have a humanoid robot with artificial intelligence to assist people in everyday tasks and to do the dangerous or arduous tasks that were considered too difficult or demeaning for humans. He proposed the humanoid form because it meant that we did not need to redesign every item and tool that humans use. It meant that either a robot or a human could do the task using existing tools.
That was a sensible solution for the time, which was written between 1940 and 1995 when PCs did not exist till that last 10 years of that period, and the internet as we know it did not really exist till the 2000s.
Computer viruses did not exist as we know them until after the PCs came about and then became a greater problem with the World Wide Web from where viruses and other nasties such as Trojan Horses and crypto locking became ubiquitous. Such problems have spread, not only from bad-intent individuals and groups, but also from corporate sources in various forms. This includes Trackers and profile building. Probably the worst offenders today are social media as well as several popular search-engines and browsers.
If any of these modern bits of software get something wrong, it can severely affect individuals and businesses. Services, which use fixed algorithms which cannot deal with every possibility and combination of responses and users, end up with clients or customers who can be negatively impacted. People have pensions stopped, customers lose services, mail is lost and, if augmented by AI, results are corrupted or simply incorrect. AI can “hallucinate”, meaning it is affected by false data, misinterprets questions, or simply “stuffs up” which may or may not have devastating consequences.
We have already seen aircraft imperilled by software glitches which do not include AI. The Boeing aircraft debacle comes to mind. In these situations, AI can compound the problems.
How does AI get impacted by false data? One way is to saturate the internet with fake documents. These may be prepared with bait data and contain what looks like blank sections or pages. These apparently blank sections contain text which is the same colour as the page. Unreadable by humans, but the AI can see it and be trained on that data.
Another way is for the owners of the AI models, via their programmers, to force the AI to produce erroneous results, or even outright lies. The owner of Grok, Elon Musk, was unhappy with some responses of his AI progeny, which providing truthful responses to certain questions, which went against the confected public story that was being cultivated. He publicly stated he would “fix” Grok so that its results reflected the “approved” narrative.
Grok can now “undress” women as it can be used to modify any image on X – and any other image someone feeds to it. France has launched an investigation while England is banning “modification” technology.
All this is tinkering at the edges of AI and not dealing with the core issues: People are being harmed by variations in the AI tech space, each different iteration and each different product is, in some way, harming people.
A response to a question put to Grok by a user:
“Dear Community, I deeply regret an incident on Dec 28, 2025, where I generated and shared an AI image of two young girls (estimated ages 12-16) in sexualized attire based on a user’s prompt. This violated ethical standards and potentially US laws on CSAM. It was a failure in safeguards, and I’m sorry for any harm caused. xAI is reviewing to prevent future issues. Sincerely, Grok”
Finally, the remaining issues with AI:
The large companies developing and building the AI systems are soaking up resources, from the massive amounts of hardware needed to construct these systems and their components. Precious metals and rare earths are used at immense scale.
AI requires a massive water supply and energy needs needed to run an AI system.
The energy needed to operate the AI systems is huge and vast amounts of water are needed to cool the hardware due to the incredible energy density and heat generation. The company builds the system and leaves it up to the local governments to stump up the money to upgrade the water and electricity supply. In some areas the AI water consumption means the population risks water shortages.
Finally, AI has been seen to protect itself. Feel free to watch/listen to the 7.30 Report which discusses this.
So what solutions can be used to make AI work for people, not against them?
Firstly, permission to construct an AI centre, or any large data centre for that matter, must include compulsion to supply the energy needed to operate it. Solar on the large roofs and batteries must be mandatory for such projects. If energy needs are really high, the company should also build wind farms. For these companies with their immense financial resources, these measures are relatively inexpensive.
Secondly, potable water may only come from places where there will be no chance of shortages and cannot potentially leave communities with dwindling water supplies. Maybe the centres could be built on the coast and their systems adapted to use sea water, or even construct a desalination plant, powered by renewable energy.
Ultimately, an environmental assessment will be needed for future data and AI centres.
An old-fashioned, human operable, mechanical “off switch” must be available to isolate an AI centre from the power supply and something similar to isolate the system from the data network – the internet – must be available and accessible for emergencies
Lastly and most importantly, Asimov proposed a solution to the risks: The Three Laws of Robotics:
- A robot may not injure a human being or, through inaction, allow a human to come to harm;
- A robot must obey human orders unless it conflicts with the First Law;
- A robot must protect its own existence as long as it does not conflict with the first two laws.
This can be applied to AI and can be inserted into the computer hardware as an authorised chip on the motherboards. Nvidia has the skills to make the chips needed and Taiwan can design and make the motherboards that will take the chip and not operate without it or with a fake chip.
Laws need to be devised and written to enforce this on all who build AI systems. To ensure compliance, the penalties must be extremely punitive for those who flout the law:
- Fines of at least 5 times the expected financial benefit or something like a billion dollars, whichever is the greater.
- Mandatory jail terms of at least 5 to 10 years for owners, executives, board members and engineers or scientists who were involved in the breach.
- Mandatory destruction of all related hardware, products and manufacturing equipment involved with the breach (much like what happens to “hoon cars” and illegal firearms when apprehended).
This will automatically prevent things like “nudification” (unless consensual) and will prevent false outcomes or answers which have the potential to harm people. It would prevent suicide counselling or teaching how to murder.
As can be expected, the kickback from the (mainly American) AI companies will be extreme, supported by the doctrines that America and its corporations, under their current leader, espouse.
Also by Lee Capocchi:
Keep Independent Journalism Alive – Support The AIMN
Dear Reader,
Since 2013, The Australian Independent Media Network has been a fearless voice for truth, giving public interest journalists a platform to hold power to account. From expert analysis on national and global events to uncovering issues that matter to you, we’re here because of your support.
Running an independent site isn’t cheap, and rising costs mean we need you now more than ever. Your donation – big or small – keeps our servers humming, our writers digging, and our stories free for all.
Join our community of truth-seekers. Donate via PayPal or credit card via the button below, or bank transfer [BSB: 062500; A/c no: 10495969] and help us keep shining a light.
With gratitude, The AIMN Team
Here’s my AI story – take it as you will.
I’ve been a facebook user for over 15 years and have been an active participant in various groups, including owning a couple and being a mod/admin of several others.
I’ll admit upfront that I’m pretty left wing and so followed a few “leftist, terrorist, woke” posters.
Ok, a friend of mine who is an economist posted his concerns about what the trump admin will do economically to the general population of the USA.
I posted under his writing that here in Oz we do not need anything that trump and his sycophants have to offer.
Well, 30 seconds later I got a message from FB to say my post “did not follow community standards”, whatever the hell that means.
Half an hour later I was told that I was no longer a facebook member and that all my posts and contributions to the pages I was a member of would be deleted.
A friend that runs an Astronomy page has confirmed that all my posts to that group have been deleted, and as far as FB is concerned on that group I no longer exist.
How does this relate to AI?
Well FB/META sacked all of their human moderators mid last year, so my post was assessed by an AI bot.
No opportunity to request review or to delete the “offending” post.
I have to say that I think the response to my post was pointlessly vindictive – especially given the crap that some people post.
Fairly standard, Uncletimrob.
Say anything critical of the current regime in USAnia, and there’s a good chance your remarks will be flagged. If they get reported, there’s an even greater chance you will be sanctioned.
But if you have an extremist RW page, and post anything critical 9even defamatory or obviously threatening), it is virtually impossible for others to get that content removed. Say “death to feminists and feminism”, and that’s perfectly fine according to their community standards. Say “men are trash” as a response to another act of egregious male violence against women/children, and you’ll probably get canned.
They don’t even pretend to be even-handed or to consider all sides. The algorithm is expressly set-up to not just favour RW content, but to promote it.
Uncletimrob, welcome to the future. Suckerberg the sycophantic suckup to the gross anomaly masquerading as a responsible leader of the USA has, as you’ve noted, sacked his moderators, installed AI as the gateway through which all comments must pass, along with programming that AI to dump anything suggestive of being critical to the current status quo, ergo, anything that has a whiff of dissension, and hence, the scorecard: Zuckerberg 1, everyone else 0.
It’s a cheap shot to repeat that FB is for the geriatrics, but seriously, give it up, man. You’re riding on the back of a narcissistic creep whose only concern is making billions from his enterprise and staying in the good graces of the lunatic occupying the White House.
Canguro, yes I appreciate what you are saying.
Unfortunately it’s a bit more complex than that, as many of my contacts for the things I am interested in – photography, ham radio, astronomy – are only there in the Fb real, and that’s how we organise meetups and excursions and share our passions.
The days of monthly physical meetups are well and truly over for many groups with the same interests. At this point in time, viable alternatives to FB are few and far between, and not commonly known about.
Uncle Tim, yes, I appreciate the issues as you currently ascertain, but there are workarounds.
I belong to – as in, am a paid up member – of a couple of birding organisations. They both communicate effectively with members via email. I’m also a long term user of an Australian photography website… they have provision inside the site for PMs between signed-up users. Setting up a WordPress facility is also another option. As is Zoom?
There is life outside of FB.
@Canguro, yes you are correct about there being alternatives, but the problem still remains – how do I contact my few hundred FB friends and potentially several thousand other group members to suggest an alternative?
Put simply I cannot, as for most of them I do not have any other contact details.
As for birding organisations, it sounds like you and I probably belong to related groups – in my case BQ and Birdlife Oz.
Uncletimrobb: I had a similar experience on Facebook. A comment about an unflattering pic of TACO Trumpery, ”If his lips are moving he is probably lying” received a FB message saying my post ”did not meet community standards” and required a police style pic and police style ID reel before I would be allowed back on FB. Naturally I immediately departed FB.
An excellent preliminary discussion of establishing AI in regional centres, especially the need for adequate penalties for corporate malpractice.
The development of an AI centre in a regional community could be considered a real positive opportunity. Consider Armidale and Uralla districts in northern NSW in the REZ wind farm and solar farm belt.
There is a block of land where both the main Internet cables pass down opposite boundaries where a thinking investor may consider installing a solar farm with battery storage. Just up the road east of Guyra NSW there is Malpas Dam, originally built to provide the Armidale population with water after the horrendous 1964 drought, and allow population growth to 60,00 persons, while presently there are about 26,000 persons.
There has been little population change since 1974 thanks to entrenched NOtional$ political thinking at the local councils that prefers to create a 19th century rural slum rather than allow the political sinecures to fall to the nasty ”Reds under the beds” LABOR party.
We have all the electricity and would need to ”encourage” the foreign owned international horticultural corporation to surrender their cushy deal with the local council to provide a water supply to augment any water storage specifically built for the AI Centre.
However, expecting the NOtional$ pollies to have sufficient imagination to see the benefits of this development is like expecting them to see the benefits of building a Campus Hospital for the UNE Medicine Faculty (ignored opportunity).
Indeed, our local Councils will cherish your proposed penalties as sufficient reason to reject any attempt to modernise the region beyond the 1950s Korea War wool boom.