5 ways AI is transforming the workplace

What Do We Mean By AI?

General AI is the stuff of movies, generally embodied in a robot or android. Think C3PO in Star Wars, Data in Star Trek, or Ava in Ex Machina. It can learn any intellectual task that a human can. Definitions of general AI get tied into philosophical knots. What is intelligence in humans? Can a machine be said to think? General AI does not yet exist, and some experts claim it never will!

Narrow AI, however, is already with us. While traditional software achieves defined goals by following steps prescribed by a programmer, narrow AI achieves those goals by mimicking human intelligence. Here are five examples of how narrow AI is transforming modern workplaces.

‍

1. Natural Language (Written)

We are a long way from having a natural sounding conversation with AI. However, using natural language for a written conversation about a given topic is achievable. Take a written manual for a production line, or a detailed description of supply chains for a logistics operation. Traditional software would allow you to search for specific text – but unless you know how something is worded, you might not find the information you’re looking for. A natural language AI system can deconstruct and analyse the written information, and then respond to typed questions. For example, if an engineer is trying to trouble-shoot a problem with machinery, rather than spending hours wading through a manual, a question typed into an app could result in some of the best considered suggestions.

2. Natural Language (Spoken)

Bots that respond to written language have been around for a while, particularly as a replacement for telephone call handlers in customer service. Responding to spoken questions is evolving more slowly. It works best where the AI can anticipate the words the human is likely to use. A worker who needs their hands free to use tools, or can’t use a keyboard because of protective gloves, can use a voice command to access information or to provide simple instructions to machinery.

3. Decision Support

You might already have been at the receiving end of AI-driven decision support when you shop online. The first thing you look at isn’t quite right, so you are provided with suggestions of other items you might want to buy. Although online shopping suggestions are sometimes perceived as bizarre, with better information, decision support systems in the workplace can help engineers to diagnose faults in equipment, clients to choose the best contractor for a job, or risk assessors to find suitable controls for identified hazards.

4. Machine Learning

Traditional software could only be enhanced by programmers modifying lines of code. AI can learn directly from the users, or even from its own efforts. The game-playing AI system AlphaZero was given only the basic rules of chess, and performed at champion level after a few hours by playing against itself multiple times. Machine learning can enhance other forms of AI. Systems based on natural language can learn to respond better to future questions. A health and safety manager who uses a vision system to understand risk and hazards can inform the AI as to whether or not the recorded incident is accurate. The AI can evaluate its "thinking" processes to see how it can enhance the observations of other events, as well as improving outcomes for the same event.

5. Vision

For more than fifty years, technologists have attempted to teach computers to recognise three-dimensional shapes from two-dimension drawings. Only in recent years, with increased processing power and advances in machine learning, has image processing developed to the point where objects can be recognised and identified in real-world images. AI vision systems used to screen mammograms have been shown to detect cancers that doctors missed, flagging up suspect areas for a human to assess. Businesses using distancing as a means of protecting staff from infection during the COVID-19 pandemic could use vision systems to flag when people stood too close to each other. Future vision systems will get better at detecting physical hazards in the workplace, as well as hazardous behaviours.

‍Conclusion

AI will not solve all our problems, but it is already being used successfully in multiple industries. In the technology race safety and health has often been left behind production, HR, procurement, sales and marketing. While they have sophisticated databases, alerts and dashboards, many EHS professionals still rely on spreadsheets and text documents. AI has so much to offer EHS – this is a race we mustn’t get left behind on. To learn more about how Protex AI is using camera software to detect risk before an accident can occur, encouraging businesses to embrace a proactive safety culture, chat to one of our product experts here 👈🏼

‍