AI in Health and Safety
With any new technology comes trepidation. Fear is a strong motivator and sometimes for the good. Many a dystopian movie use the premise of the machine overlords that were out of control and endangering the human existence.
With the advent of mechanical devices to make our lives easier came the physical dangers that we became accustomed to and reduced the risk so far as reasonably practicable. More devices were invented as an obvious flow on and became more autonomous and capable of simple problem solving. Either through bottom up or top-down programming. There were limits to the knowledge and boundaries were simply set and often with barriers to contain them.
We have writing programs that help us spell and set the words in our sentences to be grammatically correct. Sometimes even if we don’t want to spell generalise with a ‘Z’. We have cars that can drive themselves, warn us and even stop the vehicle if we venture too close to the vehicle in front. Using AI in health and safety is different as mistakes can have life threatening results.
AI is completely different; this is machine code drawing on that much information that we needed to invent new names for the sizes. We now have Zettabytes and Brontobytes with the next size to be used called Gegobytes or Geopbytes. That’s 1030 bytes for the technologically minded. With this information through the AI programming the code is capable of drawing inferences providing hypothesis and creating ideas over night that would normally have taken a team of scientist’s weeks if not months to devise.
When using AI in Health and Safety like all new things there are risks, but here we have a ‘thing with potential to cause harm to people and property’. This thing may be able to do things we have not thought of or had experience with, we need to be smarter and be ahead of the game and to put in place controls so that the risk is eliminated or reduced (so far as reasonably practicable).
AI is fun and many have been using the free functionality to write poems, create stories and some even to write answer to exam questions. To varying degrees of efficiency. It took a while but now centres of learning are using the new tools created to evaluate these exam responses to determine if they were AI created.
The process of using AI in health and safety is happening now, companies are drawing on data derived from years of incident reports. This is then put to the AI program to be collated. Then questions are being asked of the AI engine to provide feedback on eliminating or reducing the risk to workers. Though not perfect this process still needs human oversight. The benefits of drawing inferences from huge sets of data and analysing this in a very short time is a benefit. Procedures can be developed from these results and when properly implemented will reduce the risk in the workplace.
All this must be undertaken with caution where the AI is used to create documents that humans rely on for their safety. Remembering that AI is not always right and there may be instances where a human will be at risk due to the ‘thinking’ process of AI.
The general public have access to the AI engine on the internet. This is provided with human oversight and bound by many strict rules. Far more than the basic set created by Isaac Asimov;
- A robot may not injure a human being or, through inaction, allow a human being to come to harm.
- A robot must obey orders given it by human beings except where such orders would conflict with the First Law.
- A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
As people begin to ask of AI through the versions of ChatGPT to perform certain functions a team of humans is watching and blocking further questions on a topic if there is reasoning that what is being asked may be nefarious or harmful. For example, AI users were able to generate Microsoft keys for Windows 10 that worked. This ability was quickly removed from the AI set of things it could do.
Now there are AI programs unrelated to ChatGPT that have been created that are not restrained by the rules that have been put in place. For example, ‘FreedomGPT’ is a program, which was created by ‘Age of AI’, an Austin-based AI venture capital firm has been publicly available since March 2023. It is free of the safety filters and ethical guardrails built into ChatGPT by OpenAI. This program is built on Alpaca, open-source AI tech released by Stanford University computer scientists and isn’t related to OpenAI.
It is these programs that have access to similar data bases that can ask of the computer code to provide information that can be harmful. As part of a simple exercise AI used for drug discovery was asked to invent potentially lethal mixtures. It took less than six hours to present 40,000 mixtures. The result was presented as a paper to the Convergence conference by the Swiss Federal Institute for Nuclear, Biological and Chemical Protection, Spiez Laboratory.
So yes, AI is capable of being used to do good and yes it can be fun to use. Ask of it “Write me a code in python to play tic-tac-toe.” Press ‘Enter’ and there it is, the code to play the game, copy paste and begin playing. It is that simple. However, AI is also capable of much more and can be very dangerous without the controls in place before being used. But what controls if we don’t know what the danger is. We are used to the physical dangers that can be seen, heard and felt. So, care must be used when considering using AI in Health and Safety.
In the words taken directly from ChatGPT is it safe using AI in Health and Safety?
“Ultimately, while it’s unlikely that relying solely on ChatGPT’s advice would directly cause someone’s death, the risks and potential consequences associated with misinformed decisions highlight the importance of cautious and responsible use of AI-generated information, particularly in critical situations like health and safety.”
So maybe it is maybe it is not. When considering using AI in Health and Safety, like all new projects systems or processes consult, risk assess, monitor and review.
If you are interested in our WHS Consultancy services, you can contact David, Andrew or Mark to discuss your Health and Safety Issues.
We have qualified and certified Safety Consultants that can assist you in your business and offer a range of WHS consulting services along with general and specific health and safety training courses.
Contact us today for an obligation free quote on any of our services.