The point about artificial intelligence raises several questions about its impact on the broader economy: which jobs will be most affected by the technology, how will these changes occur, and how will these changes be felt?
Senior Lecturer in Robotics, University of Sheffield
BT recently announced that it will reduce its workforce by 55,000, with around 11,000 of those related to the use of artificial intelligence (AI).
The point about artificial intelligence raises several questions about its impact on the broader economy: which jobs will be most affected by the technology, how will these changes occur, and how will these changes be felt?
The development of technology and its accompanying effect on job security has been a recurring theme since the industrial revolution. If once mechanization was the cause of anxiety about job losses, today it is more able to use artificial intelligence algorithms. But for many job categories the employment of humans will remain essential for the foreseeable future.
The technology behind the current revolution is mainly what are known as large language models (LLM), which are able to understand questions and generate answers in natural language. It is the basis for OpenAI's ChatGPT, Google's Bard system, and Microsoft's Bing AI (which itself uses ChatGPT).
All of these are neural networks: mathematical calculation systems that are roughly modeled in the way nerve cells (neurons) work in the human brain. These complex neural networks are trained on text - often sourced from the Internet.
The training process allows the user to ask a question in conversational language and for the algorithm to break the question down into components. These components are then processed to generate a response appropriate to the question asked.
The result is a system that is able to provide logical answers to almost any question that is asked. The implications are wider than it seems.
Humans in the loop
In the same way that GPS navigation can replace the driver's need to know the route, AI provides an opportunity for employees to get all the information they need at their fingertips, without "googling".
In effect, it takes humans out of the loop, meaning any situation where a human's job involves searching for items and making connections between them could be at risk. The most prominent example here is call center jobs.
However, there is still a possibility that the public will not agree that artificial intelligence will solve their problems, even if the waiting times for calls are much shorter.
For any manual work, the risk of being replaced by artificial intelligence is very remote. While robotics are becoming more capable and skilled, they operate in very limited environments. It relies on sensors that provide information about the world and then make decisions based on that imperfect data.
Artificial intelligence is not yet ready for this work environment, the world is a messy and unsafe place, and humans are excellent adaptors. Plumbers, electricians and complex jobs in industry – for example, automotive or aircraft – face little, if any, competition in the short to medium term.
However, the real impact of AI may be felt in terms of savings and efficiency rather than outright job replacement. It is likely that the technology will find rapid traction as an assistant to humans. This is already happening, especially in areas such as software development.
Instead of using Google to find out how to write a certain piece of code, it's much more efficient to ask ChatGPT. The answer can be tailored solely to the user's requirements, given efficiently and without unnecessary detail.
critical safety systems
This type of application will become more common as future AI tools become true intelligent assistants. The question of whether companies use this as an excuse to reduce workforce becomes dependent on their workload.
As the UK suffers from a shortage of STEM (Science, Technology, Engineering and Maths) graduates, particularly in fields such as engineering, it is unlikely that there will be any job losses in this field, just a more efficient way of dealing with the current workload.
The conclusion relies on the team making the most of the opportunities that technology provides. Naturally, there will always be skepticism, and the adoption of AI in the development of safety-critical systems, such as medicine, will take a long time. The reason for this is that trust is required in the developer and the simplest way in which technology develops is by placing a person at the heart of the process.
This is a critical issue because these LLM models are trained over the internet, so biases and mistakes creep in. Worse, they may also arise from malicious intent, which allows incorrect or even deliberately misleading training data to be presented.
Cyber security becomes a growing concern as systems become more networked, as does the source of data used to build AI. LLMs rely on open information as a building block that refines it through interaction. This raises the possibility of new methods of attacking systems by creating deliberate lies.
For example, hackers can create malicious websites and put them in places where they are likely to be will be collected by an AI chatbot. Because of the requirement to train the systems on a lot of data, it is difficult to make sure that they are all correct.
This means that as workers, we must seek to harness the capabilities of AI systems and use them to their full potential. This means always questioning what we get from them, instead of just blindly trusting their output. This period is reminiscent of the early days of GPS, when the systems often led users on roads unsuitable for their vehicles.
If we apply a skeptical mindset to how we use this new tool, we will maximize its capabilities while simultaneously increasing the workforce - as we have seen in all previous industrial revolutions.
One response
https://google.wixsite.com/meir-ai/post/%D7%9E%D7%90%D7%9E%D7%A8%D7%99%D7%9D-1