BT recently announced that it will cut 55,000 jobs, about 11,000 of which will be related to the use of artificial intelligence (AI). The remaining reductions were due to business efficiencies, such as replacing copper cables with more reliable fiber optics.
Allegations about AI raise some questions about its impact on the broader economy. Which jobs will be most impacted by technology, how will that change happen, and what will that change feel like?
The development of technology and its associated impact on job security has been a recurring theme since the industrial revolution. Where once mechanization was the cause of unemployment fears, it is now being driven by smarter AI algorithms. However, for many or most occupations, human employment will continue to be important for the foreseeable future. The technology behind this current revolution, largely known as large-scale language models (LLMs), can generate relatively human-like responses to questions. It is the foundation of OpenAI’s ChatGPT, Google’s Bard system, and Microsoft’s Bing AI.
These are all neural networks. A mathematical computing system that loosely models how nerve cells (neurons) fire in the human brain. These complex neural networks are trained or familiarized with text, often from the Internet.
The training process allows the user to ask questions in conversational language and allows the algorithm to decompose the question into its components. These components are processed to generate appropriate responses to the questions asked.
The result is a system that can provide intelligent answers to any question. Its impact is far-reaching than you might think.
human in loop
In the same way that GPS navigation for drivers can replace the need to know a route, AI gives employees the opportunity to have all the information they need at their fingertips without having to “Google” it. .
You’re effectively taking the human out of the loop. In other words, any situation in which a human labors to search for items or create links between items can be compromised. The most obvious example here is call center work.
But even if phone wait times drop significantly, the public may still be reluctant to accept that AI will solve their problems.
Manual work has little risk of replacement. Robotics are becoming more capable and dexterous, but they operate in highly constrained environments. We rely on sensors to inform us about the world and to make decisions based on this imperfect data.
AI is not yet ready for this work space. The world is a messy and uncertain place that adaptive humans are good at. Complex jobs in manufacturing, such as plumbers, electricians, and automobiles and aircraft, face little or no competition in the long run. semester.
However, the true impact of AI may be felt in terms of efficiency savings rather than replacing jobs entirely. This technology could quickly become a human assistant. This is already happening, especially in areas such as software development.
Asking ChatGPT is much more efficient than using Google to find out how to write a particular piece of code. The returned solutions are tailored precisely to the user’s requirements and efficiently delivered without unnecessary details.
safety critical system
This type of application will become even more common as future AI tools become truly intelligent assistants. Whether or not companies use this as an excuse to consider layoffs depends on their workload.
The UK suffers from a shortage of Stem (Science, Technology, Engineering and Mathematics) graduates, especially in fields such as engineering, so jobs in this sector are unlikely to be lost and there is a more efficient way to deal with the status quo. just a way. amount of work.
It all depends on your staff being able to make the most of the opportunities presented by technology. Of course, there will always be skepticism, and it will take a long time to introduce AI into the development of safety-critical systems such as healthcare. This is because trust in developers is key and the easiest way to develop is to put humans at the center of the process.
This is important because these LLMs are trained using the internet, so biases and errors are factored in. These can happen, for example, by chance through a person attending a particular event simply because that person shares the same name with another person. More seriously, they can also be malicious and intentionally allow the presentation of false or intentionally misleading training data.
As systems become more networked, so too are concerns about cybersecurity, as are the sources of data used to build AI. LLM relies on open information as the building blocks that are refined through dialogue. This raises the possibility of new ways to attack systems by creating deliberate deceptions.
For example, a hacker could create a malicious site and place it in a location likely to be detected by an AI chatbot. Since the system has to be trained on a large amount of data, it is difficult to verify that everything is correct.
This means that, as workers, we should strive to leverage the capabilities of AI systems and exploit their full potential. This means always questioning what you receive from them rather than blindly trusting their accomplishments. This period is reminiscent of the early days of GPS. At that time, the system often directed users to roads that were not suitable for vehicles.
If we apply skepticism to the use of this new tool, as we have seen in all previous industrial revolutions, we can maximize its capabilities while increasing our workforce at the same time.
