Hi Dr Elise Tapping. So, AI is not yet capable of making autonomous decisions in war, but rather acts as an assistant that provides intelligence to support human decision-making. However, AI also plays a role of a scout that can shape how humans perceive the situation and the available options. This can be dangerous if AI introduces errors or biases that affect the quality of human judgment. Therefore, AI will need to be used with more responsibility and caution in war situations.

Hi AMK, great response and definitely agree, at the moment with the ethical and safety aspects, even if AI is able to be used in an autonomous way in the battlefield, it would not get pass the regulations put in place. I do however think that, yes it can support human decision making, as you have stated, but I would say that it can dramatically improve not only the accuracy of decision making, but the time in which decisions are made. 

As part of the IET AI TN, we are doing a webinar on AI in conflicted environments in November where we are discussing exactly this, where anyone can sign up here. It would be great to get some questions from people such as yourself who understands not only the utility, but the danger of having these technologies used in the battlefield.

At what point does the human "making the decision" come down to clicking on a popup message

Enemy tank detected. Destroy it? [Yes] [No]

Thats a very good point to raise, but is it similar to the daily practices we face, do you want to take the fastest route, yes/no? This is obviously simplifying the issues with the ethics and safety when taking this to the battlefield. Though this could be debated, there is still a human in the loop of the decision, so would it just be speeding up the decision that was always going to occur?  And what would be required to have the same threshold we hold between our own judgement, i.e. human error, Vs that of AI even if it's more accurate? Just food for thought!

What will happen if the GPS data seen by the human operator is hacked or purposely delayed by A1??  

This would be a great question to pose to Chat GPT.  Can AI understand where it's weakness are and try and improve them? 

I tuely believe that technology/engineering can makes our lives better.  AI is just an extension of that and as an assistant it will help enhance our lifes or job role.  Can it or will it ever replace me, well NO.  I am what I am.  I am unique.  

There are many examples of people following the lead of technology and getting stuck.  Satnav that tell people to drive into the sea and such.  At what point did the human realise that they were driving into the sea, just before they enter, when they are waist deep, or neck deep?  Conversely ask AI to help improve your CV and it can give some fantastic suggestions for you the human to review and use wholly or in part there of.

So my answer to the question above is

Can Ai make battlefield decisions.  No

Should it be allowed to make battlefield decisions.  NO

just filling in the 'missing' data and therefore the human making a better informed decision.   YES

A well trained Army/Navy/Airforce with the assistance of AI will have a significant tactical advantage over an enemy in the standard battlespace. 

Remember that the security of your AI is now entering the battlespace and some future warfare will consist off the shutdown of remote systems to gain the tactical advantage.  As a scenario consider somewhere like Sizewell or Hinckley Point being taken off line and shut down by a computer virus (other nasties are avaiable). 

Consider your AI being taken off line or shut down when you need it the most, hence the need still remains for highly trainied and experience military personnel

I completely agree! When it comes to battlefield decisions, there is a consensus that AI should not be authorized to make such determinations. In the standard battlespace, a proficient military equipped with AI support can gain a substantial tactical edge over adversaries. Nevertheless, as you astutely highlighted, there are potential hazards linked to an excessive reliance on AI. The security of AI systems is an escalating concern, as these systems are susceptible to hacking or malicious shutdowns.