Blog Archive

Saturday, February 1, 2025

The Agentic Era, My worries about AI, Quote for the Month

The Agentic Era

We are in the Agentic Era. We use agents, such as Google's AI agent Gemini, for more and more tasks.  There are more and more of them and the months go by. Big names such as OpenAI's Chat GPT was given a run for its money with the launch of the latest from the Chinese AI, Deepseek which caused a minor market correction.  Last month, Google produced a short video to show where this is going for the general public. What I like about it is that it demonstrates the use of AI in everyday tasks that most people will be able to relate to. Watch this four minute video which I think is informative:


My worries about AI


I have been reading a book called The Coming Wave by Mustafa Suleyman, the co-founder of Deepmind and one of today's most experienced AI experts. He is worried that AI could end up destroying civilisation as we know it. I have summarised some of my main concerns below.



Deepfakes everywhere - how will we know what's genuine? From fake news to fake videos, I think we are going to have to rely more on the 'established' press if we want to believe most of what we read. When we have Trump and his admirers already re-writing facts, AI's is able to do this far more powerfully than any human. Services such as the BBC's 'Verify' will become more and more vital for our understanding of the new (and our sanity).

How do you spot a deepfake?

I worry about the ability for criminals to get their hands on advanced AI that will shut down our technology and mean we can't access our money, our data, our lives. An example of this was in 2017, when the WannaCry attack shut down the NHS by exploiting a weakness in old Microsoft software. It caused absolute chaos. However, in the future, AI will make it far easier to identify weaknesses in almost any software. This worries me far more than the possibility that AI will take over people's jobs and start running the world. It won't. It will help people be far more productive in their work and will change the nature of work. Some jobs will go, others will be created. It will make us richer. Unless someone comes along and destroys all of our institutions and great companies by undermining the very software we use to exist.

This ransomware caused absolute chaos across the NHS

At the end of the last century we worried about the Millennium Bug. Now we should worry more about how criminals and evils states (like Putin) will attempt to destroy our civilisations by using AI to identify and exploit weaknesses in the systems we rely on every moment of the day. I hope that every government department and every private company takes this extremely seriously. It will take some serious international co-operation to control and regulate rogue AI. The chances are that the COVID-19 pandemic started from a Chinese lab leak. This is another area of concern where AI will boost our ability to produce new life-saving medicines. But with that comes the challenge of establishing very high lab standards to reduce the risk of accidental or intentional leaks.

Technology growth - We ain't seen nothing yet!!

It looks as if the speed with which AI transforms our world is going to surprise most people and it's likely that we won't be ready for it. Chinese AI seems to be open source which means anyone can copy it and modify it. Again I think we should be rapidly putting international measures in place to control its use and regulate for its mis-use. The key to success in this area is to work with our friends and not so friendly states to establish protocols that will protect us all.




Quote for the Month
In 2005, Vladimir Putin visited the 60th anniversary commemoration at Auschwitz, giving a speech in which he said it was “inconceivable to think that people are capable of such barbarity”
Hypocrite of the Century award goes to the evil Vladimir Putin.

Last Monday was the 80th anniversary ceremony of the liberation of Auschwitz on Holocaust Memorial Day. Putin was not invited.






No comments:

Post a Comment