OpenAI’s ChatGPT is under stricter regulatory scrutiny – Copyright AFP Fabrice COFFRINI
Many companies, including Goldman Sachs, a saint of legitimacy and discretion, are proclaiming the end of many professions for AI. The emperors are now admiring their new clothes in mirrors and trying to make money with them.
Goldman Sachs now predicts 300 million jobs will be lost. Euronews has a long list of occupations that are being affected by AI. …and so on. You would think this all happened yesterday.
It wasn’t. It’s a shame that the superficiality of clickbait doesn’t live up to the facts.
A standard discussion of AI and jobs is:
- AI can write. Sure, it is possible and competently written. That technology originated separately years ago and was simply ported over to AI. AI has learned how to do what other technologies have already done. Whether or not you can write well enough to read in a semi-literate culture is another matter.
- AI can create graphic arts and fake images. So it can. So do much older technologies. Deepfakes predate current AI by about a decade. Guess where the AI learned how to fake it. You can manage objects, but can you do Renoir-esque composition? Improbable at best.
- AI can do math. The same goes for the abacus. Can you identify the mathematical problem or information you need rather than just come up with an answer from the metrics you get? …because that’s what nearly every other form of computational technology does. You don’t need AI, let alone a chatbot, for that.
- AI can do news analysis. Well, can it be? Humans seem to be pretty bad at analyzing news, so AI has to learn from somewhere. In this case, the AI has to learn how to analyze news from mainstream media. How “threatened” can you feel?
As you can see, AI is really just doing what other technologies have done and applying its systems. The main advantage of AI and why it’s economical is that it’s fast. Can perform assigned tasks as permitted by the technician, such as creating software, graphics, videos, etc. AI can even respond to very vague instructions in some cases.
What’s going on with Fate and Darkness at the moment is that our limited understanding of technology allows us to make astrological predictions at best. Some have not carefully predicted what this technology will actually do, but the speculation is clear.
In other words, AI critics are dealing with strings that don’t exist yet.
- AI obtains information from various sources using search. We all know how efficient search is these days. These searches epitomize laziness and are often worse than they were in the 90s.
- They are looking for something that is often written to commercial SEO and SEM specifications. A benchmark level of 5% of all data inputs is incorrect in some way. The quality of information is often horrible and outdated. What exactly is this a trusted foundation for?
- AI is completely dependent on other technologies to function. That’s hardly an advantage. that’s a weakness. We can also use these same technologies to deconstruct what the AI is doing and see how it did things. You can AI out AI by simply adding a variant to the original.
- A chatbot interface is a very useful feature, probably more useful than anything else at the moment. A quick response saves time.
- The AI works according to assigned tasks based on instructions from the interface.Do not initiate or invent tasks. AI assigned by an idiot, or likely an idiot, management team can produce silly things.
Now, a more exotic business-level problem:
How do you fit AI into physical business systems and workflows? why? to achieve what? what is the value? These are not rhetorical questions. You can spend millions of dollars on it whether you need it or not.
Will it work on blockchain? It will probably have to do it. This is one of the strengths of AI, but try applying it to the NYSE and other high volume trades.
How secure is AI? Could it be hacked and abused? Probably. Have we heard a word about it? not much.
Who are the business drivers and decision makers? people. Not AI. There’s a reason for that.
Can AI manage the “dashboard” business? yes. can it be sold? Only if it’s very well scripted can the scripting process become clumsy and very competitive. The combination of clumsiness and competitive issues is less of a mystery.
Markets, business performance, or anything else called “analysis” requires a very strong and often specialized knowledge base. Can AI do it? are you sure It can get messy if you’re not sure.
Read the Euronews link to see which jobs will be impacted by the expected impact of AI. Now let’s consider the actual reality of these jobs and what can go wrong.
Next, have the AI write a report. read the report. Why don’t you put your name in it? Are they delivering the important information they are supposed to be delivering? Are reports actually useful or are they shopping lists?
Arts and Media – How “creative” is your AI? Does it produce hurtful prose or devastating imagery? How competitive is it?
Remember – all of these will be retired next year at the latest. There will be an AI bubble in investing. There will be an AI bubble in consumer use. Want to put billions of dollars into it? maybe not.
At this point AI is a fetus. When it can act independently, it becomes a true intelligence. Until then it’s scripted and subject to environment limitations. AI is not truly “intelligent” unless it is used intelligently.
Hope is based on positive thinking. something good will happen Can AI make it happen?
The opinions expressed in this article are those of the author. They do not reflect the opinions or views of the Digital Journal or its members.