A fallen down tree in the Jura

On Being Told Not to Fear AI

Reading Time: 3 minutes

They love to say “Don’t fear AI”. They all say “AI is meant to help people with their work”, “AI is meant to be a personal assistant” and more. It’s not AI we fear. It is stakeholders, governing boards and accountants. They are the people we fear. Throw in Inhuman resources for good measure. We “fear” AI because they, all the groups mentioned above, want to use AI as an excuse to fire human beings because computers are theoretically, and only theoretically, cheaper. 

AI Isn’t Cheap

It turns out that AI requires a lot of computing power, and computing power requires buildings, passive cooling, solar panels and more. It also requires network resources and a trainer, to teach the AI to do what it’s meant to do. I listened to a podcast about an AI Dungeons and Dragons that requires enormous amounts of computing power. AI Dungeons and Dragons requires a lot more than a top spec personal computer. It requires an array of computers. Is that progress?

Learning Prompt Engineering

When I was experimenting with building apps with AI I saw the limitations of AI. AI is great, but learning the art of “prompt engineering” is key. What is prompt engineering? It’s the art of learning to give instructions to an AI system so that it gives you the results you desire, at the quality you desire. It’s about iterating through prompts, until you get the AI to do precisely what you want. 

Keeping Things Simple

A huge effort is being made to get AI to respond to humans as if it was intelligent, but that intelligence, to some degree is an illusion, because it has to learn from millions of documents, over a period of time, and people continually have to give it feedback, to improve it. 

Reinforcement Learning

This is an interesting video about how AI learns. 

And Finally

As a media asset manager you learn to organise information so that it is easy to find. To a large degree that’s what AI does. It gathers a huge amount of data, and then, over time, we teach it to give us the results we expect when we ask one question or another. 

What this means is that we can spend billions on teaching AI to give us useful information, or we could spend time on writing more useful documentation, where the answers we seek are easier to find. We don’t need AI to replace humans. We need humans to organise information more intuitively. There’s a reason you come across Google Fu as a term. 

As I invest time in learning Hugo I sometimes find myself asking chatGPT questions, when I get stuck, because I can’t find the documentation I want. Sometimes I get useful answers, but at other times it is unproductive. 

The Real Value of AI

For me the real value of AI is not in helping us ask questions and get answers, it is in automating tasks that require humans a lot of time. Transcription is a useful feature of AI because it can automatically transcribe text, and then a human can double check that it is correct. AI can then automatically translate the transcription, and a human can go over the translation, and tweak it, and those tweaks can be fed back into the AI. 

With Media Asset Management AI is useful too. DV video tapes were 63 minutes long, and 63 minutes takes a long time for a human to watch, keyframe, keyword and more. AI could do the same thing within minutes, and then a human can qualify the result and tweak it where needed. 

AI should also be used for proofreading, rather than writing. AI can be taught all grammatical rules, and ensure that what is written makes sense. It could even be trained to recognise where people reading from a teleprompter would stumble, and other suggestions to make the text flow. 

Conclusion

In a rational world AI is just a tool that people can use to streamline their workflow, catch errors, speed up tasks like transcription and more. The drawback is that it gives an excuse for accountants, governing boards, stockholders and others an excuse to say “cut out the humans, let AI do the work.” Plenty of tech companies have done just that. It’s not that people fear AI, but that people don’t trust those in charge of budgets not to make immoral and irrational decisions, by firing the humans, that generate value in the first place. 

Similar Posts

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.