comment on this article

Is ‘general AI’ a pipe dream?

Russell James, Vice President of AI & Compute Strategy, Imagination Technologies, gives New Electronics his thoughts on whether 'general AI' is a pipe dream.

When we talk about ‘General AI’ what we mean is when machines are capable of understanding or learning complex contextual tasks that are equivalent to the capabilities of human beings.

According to Russel James, whether this is possible is open to question, but if it can be achieved there are a number of steps that need to be taken.

“We’re very early on the technology development curve for narrow AI, let alone general AI or strong AI. Can it be achieved, maybe, but most likely not in our lifetimes. The complexity and scale of the human brain is staggering and to achieve artificial general intelligence, this complexity needs to be largely emulated in a machine. One of the main issues is scaling up the number of neuron models and their connectivity. The average human brain has approximately 86 billion neurons and so far the largest artificial neuron structures number in the millions.”

Looking around in 2020, what are the most developed examples of AI – in business?

“Surveillance, natural language processing and targeted advertising are the best examples. Surveillance was an early adopter of narrow AI solutions to reduce the dependency of eyeballs on screens. The applications such as object detection, recognition and tracking are well suited to the narrow AI algorithmic development over the last few years. Surveillance and camera installations is growing exponentially. NLP has seen significant growth from the hyper scalers and cloud service providers, as a way to automate interaction with billions of users.”

When it comes to everyday lives, how is AI being used?

“The best examples of AI are the ones where you don’t even realise what’s actually going on and it just works seamlessly under the hood. AI is intelligently sorting and searching photo albums on your phone, while Alexa and Siri understand your question/request/garbled comment and look to respond in the correct manner. Advanced driver-assistance systems (ADAS) systems are being used to make your car safer.”

In light of the Covid-19 pandemic, what are the best ways AI is being utilised?

“The AI of today is pattern matching and so is very useful in crunching large amounts of data. Several areas that the AI of today can be used are to analyse chest x-rays for signs of lung conditions caused by COVID-19 or speeding up the development of new vaccines. Additionally, AI is already being used to analyse the potential effectiveness of existing drugs in treating the symptoms of COVID-19. Finally, AI models are being used to predict how the virus is going to spread and the effect different measures will have over time. “

Is there a sense of AI losing its ‘lustre’ for members of the C-suite, and if so why do you think that might be?

“Despite over 60 years of research, AI as a technology is at the beginning of its development curve to maturity. We’re going through a slight downturn in the monetisation of narrow AI, as its promise has proved to be more complex to implement, than initially envisioned. The reality that even narrow AI can require completely new types of processors, means that significant investment and commitment must be made to have a reasonable chance of success. Given the current volatility of world affairs and the financial markets, the decision to invest becomes are more serious and significant.”

Is "narrow AI" more rewarding, and also less dangerous?

“Narrow AI is really the only investment and work going on in businesses right now, so yes it’s more rewarding. Narrow AI can successfully perform singular tasks well and better than existing alternatives. There are dangers in the form of unintended consequences of operation and the potential for inherent biases in training and network architectures.

“That being said, narrow AI can provide solutions to real problems that users experience, which means there is money to be made.”

What checks and balances are in place to ensure AI is developed in a responsible fashion?

“In terms of general AI, there aren’t many! We are so far from developing an emerging consciousness that any checks and balances at this point would most likely be meaningless in the future as they would be largely guesswork”

For those businesses seeking to advance AI capabilities, what, in 2020, would be three top tips, so that more pilots are further developed?

“Find the real application/solution and then focus. Wanting AI in your products or business is a fool’s errand without understanding of how, what, why and when.”

“The potential for AI as a technology is enormous, but the road will be long and expensive. But if you stick with it, the returns will be spectacular.”

Neil Tyler

Comment on this article

This material is protected by MA Business copyright See Terms and Conditions. One-off usage is permitted but bulk copying is not. For multiple copies contact the sales team.

What you think about this article:

Add your comments


Your comments/feedback may be edited prior to publishing. Not all entries will be published.
Please view our Terms and Conditions before leaving a comment.

Related Articles

NFC reference design

Panthronics, a specialist in wireless technology, has added support for the ...

NI Trend Watch 2014

This report from National Instruments summarises the latest trends in the ...

Capactive sensing

This whitepaper looks at a number of capacitive sensing applications to ...

Spurring on the IoT

A team of Stanford engineers has built a radio the size of an ant – a device so ...

World IoT Day

ByteSnap Design, a specialist in embedded systems design and development, has ...

Digital consciousness

​Would you consider uploading your brain to the cloud if it meant you could ...

End game

How are IoT technologies keeping vaccines safe, in storage and transit, and in ...