Tag Archives: Adam Scott

That is what I see, there are two settings. The first one was not new, it was three weeks old when I saw (at https://www.wired.com/story/a-court-has-ruled-that-google-is-liable-for-false-statements-generated-by-ai-overviews/) ‘A Court Has Ruled That Google Is Liable for False Statements Generated by AI Overviews’ it is not entirely undeserved, but it also sets Google up for people fleecing them, so some will ‘cater’ to the need of supporting a setting that set Google up for a trap. We cannot see this directly from “Germany has issued a ruling that could reshape the operation of search engines and artificial-intelligence-based chatbots worldwide. The Munich Regional Court preliminarily ruled that Google is liable for a series of false statements generated by its AI Overviews feature, requiring the company to prevent the dissemination of erroneous or inaccurate claims through its search engine.” So, whilst some will cater to the need of false feeing that search engine, we are left with a more than slightly vulnerable Google. Whilst we see (at https://www.rmit.edu.au/about/schools-colleges/media-and-communication/industry/rmit-information-integrity-hub/the-repost/june-2026) ‘Should AI be liable for its mistakes? A German court says yes.’ Where we see “Jeannie Paterson, a law professor and co-director of the Centre for AI and Digital Ethics at the University of Melbourne, said the decision was “potentially very important” and could have ramifications for Australian consumers. The decision hinged on who is responsible for the content of AI search results. The law has traditionally considered social media platforms and search engines to be mere conduits for information, Professor Paterson told The Repost, meaning companies “only become liable if they knowingly participate” in sharing information that proves to be wrong.” I personally believe that Professor Paterson is setting up loaded dice. You see, in the first AI does not yet exist. And the second part is “who is responsible for the content of AI search results”, that answer has two stages. The first is the programmer who ‘created’ the analytical setting of predictive analytics, because that is part of any DML/LLM setting. It is not AI. And that data is also a side, because there is a massive failure of validation and verification. We see it all the time and whilst some are ‘whisking’ it away through ‘hallucinations’ I have seen the Grok side of things on data that I created and it take all without any reference other stories I had written, as such we see a programming failure. And through that the stage of “who is responsible” gets a new life and makes the water pretty murky.

That is what I see. And anyone saying I am wrong can take a long walk of a short pier. As I saw that, another stage was handed to me.

Last week we were given ‘Ford rehires human engineers after AI fails to match quality checks’ (source: BBC), this is not new, I saw this coming a mile away and I present (as pseudo evidence) ‘Is it more than buggy?’ (at https://lawlordtobe.com/2024/01/05/is-it-more-than-buggy/) and I wrote that story in January 2024 (over 2.5 years ago), as such it should count as evidence and I gave the clear settings of “On May 27th 2023 the BBC reported (at https://www.bbc.com/news/world-us-canada-65735769) that Peter LoDuca, the lawyer for the plaintiff got his material from a colleague of his at the same law firm. They relied on ChatGPT to get the brief ready.” Which now intersects with the AFR (at https://www.afr.com/work-and-careers/workplace/ai-use-in-dismissal-claims-borders-on-contempt-of-court-judge-warns-20260705-p60cob) ‘AI use in dismissal claims borders on contempt of court, judge warns’ and considering that this failures car in May 2023 when the BBC reported on this, we see a larger immature failure of other branches as well. You see, that it was tried is OK, and it failed three years ago, as such others should have stopped this as soon as they came aware. These settings all intertwine, because validation and verification is all part of these failures. As I see it, they were never made. I would be in favour of a separate tier of verifying all it produces, and these sources need to be validated. As such “after one claimant’s chatbot cited large swaths of evidence that did not exist, in a case showing the technology bedevilling the workplace umpire is now hitting the courts.” So, evidence that did not exist, where have I seen that before? (Small giggle inserted afterwards). This is why I feel that my services n technical support and customer service will be needed soon enough. When Fake AI fails to this degree. It is one small step for the AI agent to tell the customer “just press the carrier online button on the right side of your device” for this to fail and when that happens a few times, these ai agents will be pushed into the land of the Dodo soon enough. And that (until there is an actual AI) with proper validated and verified data is where that agent remains. You see, it was never rocket science. Some sales person saw the DML/LLM setting and started to call it AI, but Alan Turing had some clear settings on it all and this is not it. I believe in DML/LLM solutions, I saw an amazing application for lost and found in an airport reducing days in optionally less than an hour and there are more, but it is the, not AI, no matter how sweet they mention AI, it is the trap the salespeople set up and now that the class actions are setting in all kinds of field and personally I keep a high note on ‘Unauthorised Training Data’ and ‘AI washing’, whilst an alleged Anthropic settled for $1.5 billion for using pirated books to train its Claude AI model, I see my data transgressed upon and whilst some state that this is $1.5M per work, I was transgressed over 1700 times, as such I should be a billionaire (we can all dream can’t we?) But clearly I am not on that setting yet (to be clear I just confirmed with my wallet and my wallet is moaning due to a lack of green bills of $100. 

All these factors add up and whilst some are already seeing the lack of data, the lack of verification and the lack of validation. There is an overdue stage of properly aligning the settings we should be seeing. And that is why the class actions continue and whilst some will whip them away in settlements. And whilst we wonder why it took so long (over 3 years) for law firms to see that stage, we will see a lot more, because as I see it, the law interns believe that true time savings could be made with any ChatGPT/Claude the reality is slightly different and soon these clients will set up clauses that no AI is to be used and that is the larger failure in all this. So whilst Ford saw their failings in the early age, big software firms  aligning with what they call AI Agentic solutions will soon learn the price of that failure. And this is not just Microsoft, this is likely to effect all large software vendors. As such thousands will be hired once more and some who were pushed out in a slightly disgusting way will seek any other employment, as such these ‘embracers of Fake AI’ as I tend to call them will have a new problem and employment agencies are no longer able to get any, some who used their Agentic solutions from day one. And the fallout is soon spreading all over the world. So as I have seen these markers all over 2025, I see opportunity (for myself) and other technical support people in 2026 and 2027. The question for these firms becomes, did they treat their support people proper, or were they (as the teams goes) ‘dicks in reducing their staff members’ in this I love the quote from Walter Mitty (Ben Stiller) “This thing that you do, Ted, where you come into a place and push people out, you should know those people worked really hard to build this magazine. They believed in the motto. And I get it, you’ve got your marching orders and you have to do what you have to do, but you don’t have to be such a d*ck. Put that on a plaque and hang it at your next job.” And those who loved the part Ted Hendricks (Adam Scott) played in all this, because he was so managemently will now have a much larger problem, because I am still in contact with buddies who did my job 30 years ago and we all talk. So they now are unlikely to find anyone. So whilst they are learning that all AI is Fake AI and they could wait for for 2 decades (for True AI), but their KPI based is not that long, they now have a problem. And the is all before they figured that all data required revalidation, verification and attune it to a newer system, the markets will suddenly experience the bubble setting, that according to SoftBank CEO Masayoshi Son, who called the current artificial intelligence boom a “bubble” is an insult will be forced to do an about turn on that setting, of course those investors will have faced the write off if trillions, so they are unlikely to send Masayoshi Son a Christmas card in 2028, but that is merely my view on the matter. 

What matters is that is that these evangelists and influencers screaming “AI” are about to be found out as the new evil. There is also the groups that properly set the AI field in a DML/LLM setting, and they will be OK. If they had properly prepared their customers and aligned them with what is, I reckon that they will be OK (still a personal view on the matter).

So where are we now?
As the news is giving us more and more failures, more and more about turns from larger companies. We are seeing what could become the implosion of that bubble. The problem is that is will not implode all at once (some are unable to survive that), it is more likely then not to manifest itself as group implosions. Not all at once, but (for example) 10 explosions of 10% and when they are apart enough, some of the larger player will survive. In one setting when these judges consider that this setting was going on from 2023, making the decision that all AI assessed briefs are regarded as “clear contempt of court” we see that it would become a setting of staggered failures and when the time between these events are enough apart, the write off is optionally limited, but that is me just hoping for a reduced heartache. It is unlikely to affect me, but hoping for the worst setting is just uncivil. 

There was actually more, but I am somewhat exhausted and I have written part f all this before, just browse through my blog. I am still in a setting where I want to see who used my blog for scraping and AI washing. I doubt if I will ever find evidence that holds up in court, but with a (massively delusional) $2.55 billion which was 1700 times 1.5M at stake, one might be willing to waste a few hours on this. Anyway time for men to continue a written adventure in Abu Dhabi, time for more there too.

Have a great day.