AI dangers

There is no doubt that AI will have a huge impact on the lives of everyone in the future. It will get better and better and will have impacts for good and evil.

In many ways I have been impressed with AI. For example I asked it to colourise an old black and white photo of my dad in his RAF uniform during WW2. The colourisation was excellent.

On the downside I asked Gemini to "make a picture of Roger G3XBM" as an experiment. I tried this both in "fast" and "thinking" modes.  On both attempts the person looked nothing at all like me. It created a radio shack (far more extensive than mine, HI) and a plaque with my callsign, but the person was totally wrong. I suppose it has to find a photo of me to learn from.

This is one of the dangers of AI. People may believe what they are being told. Often it will be correct. At the same time it can be very wrong! For now, it is best to treat AI with some caution knowing it can go "rogue" and give totally false results. It can only be as good as the sources it learns from.

AI voice?

As many know, I had a stroke in 2013 that left me with a totally poor voice that many find hard to understand. Before my stroke I made quite a few YouTube videos, so I have quite a few samples of what I used to sound like.

My wife has suggested I use these to create an AI voice of me as I once sounded. I have never tried AI for this, but it sounds a good idea. The first stage will be to collect some audio samples from YouTube and then recreate my old voice using an AI tool.

If you know of suitable (preferably free) tools,  please let me know as soon as possible.

AI and data centres

There is no doubt that AI consumes hugs amounts of power. We have some way to go to match the memory and processing power of the human brain. I am sure this will get better with time although we are a long way away at present. 

So far I have been very impressed with Gemini AI that renders old black and white photos in colour in just seconds.

This creates a moral dilemma: we all want the benefits of AI without its power-hungry nature.

What can you believe?

If you look on YouTube you could believe anything.

Many of the videos are now created using AI, so the only thing you can be sure of is "this could be incorrect". It is getting harder and harder to tell what is real and what is fake. 

Unless you are absolutely sure it is safer to assume everything is fake!

AI dangers

Recently I have been watching several videos on YouTube supposed to be the sayings of Albert Einstein. The trouble is we can’t be sure he ever said these words! 

It is now very easy to make videos, photos and texts with AI, The trouble is these AI generated images and words are often believed. So, disinformation is spread. 

I am sure there are lots of good things that become possible with AI where lots of data has to be processed quickly, but there are also great dangers. 

We now have to treat EVERYTHING we see with scepticism sadly. 

Decreasing human contact

Just now I was invited to check my PayPal balance. They wanted to send me a code to my phone to verify it was me. I doubt a human was involved. 

Most chat bots are computers rather than real people.  At supermarket checkouts you are encouraged to use automatic checkouts.  Most bank transactions are getting cash from an ATM. If you are lucky you may get a call centre somewhere on the planet, where a person might understand you and you might understand them! You are just a sort code and account number - no personal contact.

More and more there is less human contact. All this is done in the name of "progress" and to save money or increase profits. 

In my view we are going backwards.

AI and jobs

It seems we are facing a new industrial revolution around the world. Many of the tasks done by humans can be done faster and better by machines running AI.  Jobs we traditionally thought only humans could do are being replaced.

This is a two-edged sword. Boring repetitive work will go. Also, many skilled jobs will go. 

The good side is many people can be employed more usefully. The bad side is many traditional jobs will go and bad actors could use AI for malicious means.

In my view, these things are here to stay and we have to adapt.

The safest jobs will be those that cannot be done by machines i.e. those that need real human interaction.

Many of the jobs that seemed "good" when I was young may not be the jobs to look for in the future.

The murky world

There was a lot of panic when the Chinese launched their AI model DeepSeek earlier this year. 

The West would like to eavesdrop on countries they think could be a threat. I am sure there is a lot that goes on that most have no knowledge of. Perhaps it's best we don't know. 

One thing we can be sure about: every major power is spying!

See https://www.bbc.co.uk/news/articles/c4gez754mn6o

The fog of war - NOT amateur radio

Recently Israel and Iran have been fighting a war in the air. The Ukraine war has been raging since early 2022. Israel has been at war in Gaza since it was attacked in late 2023.

One common factor is "the fog of war". That is each side tries to outdo the other by spreading often false data. Recently, AI has been used.

All I am saying is this: it is very very easy to believe everything you see whereas it is quite common for this data to be wrong, especially during wars. Caveat emptor.

See https://www.bbc.co.uk/news/articles/c0k78715enxo .

AI and circuit design

For some time now software engineers have being using AI to generate code. It would appear that hardware designers could also use AI to design circuits. It would appear that AI is capable of doing everything as well as humans (and more?) , if trained.

The future is indeed scary. When we eventually meet intelligent life in outer space how do we know if it is conscious? 

Where will AI be on Earth in 50 years' time?

The future of AI - NOT amateur radio

To say I am amazed and frightened by AI is an understatement. 

Recently I watched a video imagining a street scene in 1897. I have used AI every day to get the daily sunspot number. I put a few placenames in AI and got back huge amounts of accurate data. A friend said he used AI for most of his software coding. 

A few years ago AI often produced rubbish and could not be trusted. Now there are far fewer occasions when it "goes rogue". Today I discovered a painting produced by a humanoid robot that sold at action for over £1M. As time goes by AI will get even better! Imagine AI in 40 years' time.  It cannot be uninvented. There is little doubt that many jobs will be done in future by AI. 

We have to adapt.

See Meet the AI robot whose artwork sold for over $1m - BBC News

Intelligent life in outer space - NOT amateur radio

One of the points which I picked up from my recent University of the Third Age course in Cambridge was about the advances in AI. 

Within a few years AI is likely to exceed the intelligence of human beings. When we scan the skies for intelligent life elsewhere we may come across civilisations (e.g. via SETI) that may not be sentient. 

What is consciousness? I think it is beings that are self-aware.  Although vast computers may have vast amounts of knowledge, this is not the same as self-awareness and consciousness. 

I can imagine some planets even slightly more advanced than ours which have AI far more advanced than ours.

AI lectures - NOT amateur radio

On Friday, I finished a 10 week course at the Cambridge University of the Third Age about AI. It was an excellent course and got me thinking.  It is generally known that AI will well exceed humans in the next few years. 

This morning I watched an AI generated video imagining life in late Victorian times. If I had not known it was AI generated, I would never have known. It was amazingly good. This is just one of the negatives of AI. It is already hard to tell what is fake and what is real. Imagine how this will be in just 10 years' time!

In many ways AI will benefit mankind. It is also very dangerous in the wrong hands. I imagine that the military in the USA, China and Russia are already working hard to counter each other. I imagine they are all working on the idea of mutually assured destruction (MAD) remembering what ever you can do to me I can do to you. It is like nuclear weapons.

Am I in favour of AI?  It is here and we have to live with it whether malign or benign.

The dangers of AI - NOT amateur radio

There is no doubt that AI can be very good at analysing lots of data very fast and can be much better than a human. Large language models are trained and GPT4 is vastly bigger and better than the last version. Mistakes still occur, yet many people believe it is true!

Recently I have seen several things on Facebook that may have been done with AI. The trouble now is telling what is real and what is not. As time goes by, AI will only get better. All I can advice is you take great care and assume anything a person is supposed to have said is AI unless you know for certain it is real.

DeepSeek AI - NOT amateur radio

There is a new AI program from China that is "upsetting the apple cart".  It is called DeepSeek and is supposed to be better than ChatGPT. When I went to try it yesterday, it was unavailable due to a malicious attack.

There is little doubt that AI will get even better over time and will have a great impact on our lives. In many ways it will be a force for good, but it will cost us jobs and could be dangerous in the wrong hands.

See https://chat.deepseek.com/ .

AI - NOT amateur radio

This morning I had an email inviting me to try Google’s new AI tool called Gemini. 

It makes some mistakes, but in the main it produces amazing text and images based on the information available in seconds.

In coming years it is likely to get even better. Used for good, it a powerful tool. In the wrong hands it could be very dangerous.

Soldersmoke AI experiment

As an experiment, Soldersmoke did an experimental AI generated podcast. It is frightening. 

If this is possible now, what will be possible in 2050? To me, it sounded incredible. You would never have known it was AI generated.

It is frightening how realistic this podcast sounds, making you wonder how AI could be used to distort truth in future e.g. in elections.

See http://www.soldersmoke.com/About SolderSmoke 2.mp3

AI key - NOT amateur radio

According to The Knowledge today, Microsoft is introducing an AI key on its keyboards . This links to their AI tool in Windows (Copilot) which is linked to ChatGPT.

AI - I HATE it!!! - NOT amateur radio

At one time you could speak with a human. 

I have just wasted 90 minutes of my life trying to ask my SIM only service provider how to transfer numbers when I upgrade my mobile phones. After 3 humans and several AI robots, I have given up. 

All I wanted to know was how? A human could have told me in 60 seconds.

Maybe I shall go somewhere else?

Artificial Intelligence (AI) - NOT amateur radio

Until this year, most had not heard of AI. Now it is widespread.  

There is no doubt that with big computers that learn we can all benefit from the huge processing power. What a human brain can do in days could be done by an AI computer in seconds. This could really help in such things as fast cancer diagnosis.

At the same time it could be a great cause of harm in the wrong hands. 

Some time ago, I did an experiment with AI. I put my name and callsign in and within seconds it produced line after line about me, most which was accurate. I then put my son's name in (he is an internationally famous jazz musician) and it came back with rubbish. The danger is that people may accept what AI tells them as absolute truth. Imagine a hostile power using AI. We can be sure the military services both east and west are "on the case".

I can see both good and bad. We have interacted with AI for years e.g. with chat bots and phone chats that, for example, ask you to "press 2 if you need to discuss future cruises".  In the future they will be far better.

At the moment, I am reserving judgement.