Is A.I. more than the sum of its parts?

[Sagittarius A-Star]

Business survey:
Almost 20% of companies in Germany already using artificial intelligence (AI) consider it easy or very easy to replace workers with (technical) degrees. To do that, they would rely on a worker without a degree who is supported by AI.

Source:
https://www.ifo.de/en/facts/2026-06...icial-intelligence-alternative-qualifications

 

[jack action]

To do that, they would rely on a worker without a degree who is supported by AI.

That is funny. I can imagine technical workers working with AI being more efficient, thus requiring less of them in the work force (or have them work less on technical issues).

But I cannot imagine the use of AI leading to unskill workers doing the job. How does one know the job is done appropriately?

Second, I find quite weird that it is considered essential to have a population with some general knowledge, such as geography, history, arts, etc. included in their degrees, but that "general" technical knowledge will become superflous stuff that nobody needs.

 

[Dale]

How does one know the job is done appropriately?

Indeed.

https://www.reuters.com/legal/litig...t-misused-ai-disqualifies-lawyers-2026-06-09/

 

[256bits]

Indeed.

https://www.reuters.com/legal/litig...t-misused-ai-disqualifies-lawyers-2026-06-09/

Reuters wants me to subscribe, but no problem as I had heard of this from another source.
In essence, it was 2 Chats arguing between themselves their side of the court case, with the briefings and the false references being produced by the LLM in separate sessions for each lawyer.

In other news,
Google may appeal their loss in court regarding search pre-ambles generated by AI. Google position is that since it comes from a user search it is really part of the search, with the the user responsible to check the veracity of the summary. Judge said, on the contrary, the pre-amble summary is new generated prose no where to be found in a reference, with Google AI being themselves responsible for errors and correctedness. Google is accountable for their imperfect product.

 

[256bits]

Second, I find quite weird that it is considered essential to have a population with some general knowledge, such as geography, history, arts, etc. included in their degrees, but that "general" technical knowledge will become superflous stuff that nobody needs.

The bate headline should have read " Of companies using AI, 80% find it difficult to impossible to replace workers."

 

[Dale]

Google AI being themselves responsible for errors and correctedness. Google is accountable for their imperfect product.

I think that is as it should be. Software engineers have used the title "engineer" for decades. This is one of the important things that engineers do. They take responsibility for the products they design.

 

[Sagittarius A-Star]

The bate headline should have read " Of companies using AI, 80% find it difficult to impossible to replace workers."

Another possible headline: " Of companies using AI, 74.7% find it possible to replace workers."

https://www.ifo.de/en/facts/2026-06...icial-intelligence-alternative-qualifications

 

[256bits]

Another possible headline: " Of companies using AI, 74.7% find it possible to replace workers."

https://www.ifo.de/en/facts/2026-06...icial-intelligence-alternative-qualifications

Of course there is that too as an interpretation of survey results.

The article by the author as written to substantiate a biased premise that companies ( surveyed ) are pro-AI, and are willing to view the results of AI output as being trustworthy enough to replace qualified workers. The premise, which may or may not be true, is that of the author, as her own interpretation of the results, without giving supporting information, such as how the survey was carried out ( example - the ratio of questionaires answered to number sent out, the size of the companies, who from the company replied - CEO, junior executive - , the ratio of companies questioned who have adopted AI to those who have not adopted AI and thus are not tallied in the results as seen in the graph, etc ).
She has committed a Type I Error in the heading and summary just by the way both have been written, and perhaps a naive company will rely upon her analysis to invest in AI, ending up with disastrous results. ( Would you take a drug that has been shown to have only 20% success. Would you invest in AI by replacing qualified employees with AI enhanced unqualified employees knowing the success is 20% ).

Secondly, the data is an invalid set for a company to rely upon for the direction any one company should take. The questionaire asks "Is it possible ... ", which is an attempt to predict the future. As written, it is unknown if the companies have or have not tried the replacement, indicating either success or failure in their response.
The author is attempting to replace 'Viewpoint', with 'Success' or lack thereof, as being a valid criteria.
Were the companies asked "Have you attempted to replace ... " upon which data could be obtained from past experience to portray pitfalls for any company contemplating AI adoption, rather than join the crowd as everyone else is doing it so we must too.