AI use in weapons a growing concern for ethical investors

Local ethical fund manager Pathfinder is among other ethical players globally grappling with the all new conundrums created by the rapid growth of AI  - the development of AI-powered weapons systems becoming the latest issue to raise a red flag.

“Who decides on whether the weapon is fired and what level of human interaction is there in that and and is it a human veto, or is it an entirely autonomous process?” asks Pathfinder co-founder and chief executive John Berry, in conversation with Good Returns.

“That there is a massive ethical issue, and there's a lot that can go wrong.”

Managers, while naturally adept at due diligence, have a fresh challenge on their hands understanding how the technology of the companies they may be invested in or considering investing in, is being used, when there is so much uncertainty and mystery surrounding the end-use of AI in all manner of applications, weapons included.

Any notion that AI-powered weapons are still a way off was dashed last year by a Harvard University study, which found militaries already developing and deploying systems built on AI.

“Who decides on whether the weapon is fired and what level of human interaction is there in that and and is it a human veto, or is it an entirely autonomous process?”

The find prompted the authors to sound several warnings suggesting the use of drone or robot weapons could make it easier for countries to go to war, raising concerns that it could see AI research censored to keep secret the technology behind the development of systems, and how AI-powered autonomous technology could allow militaries to shirk responsibility for decision-making.

The issue of copyright of source material remains a live issue, says John Berry, with all eyes on the Meta Federal court case in San Francisco in which authors are challenging the tech giant’s right to use their material to train the company’s LLM (Large Language Model).

Berry says the rise of AI does present some interesting opportunities from an ethical investment perspective.

“I think we're going to see new solutions generated by AI, if you think of AI being able to deal with really complex data sets and also find connectivity of ideas which are difficult, if not impossible, for humans to actually do manually.

“So they may be around energy challenges, solving energy challenges or resource efficiency or or just better mapping of extreme weather events and greater warnings around which would be extremely helpful at times.”

Energy use is another consideration for Pathfinder, with ChatGPT estimated to need as much as 10 times as much processing power as Google to conduct a search. It is a double-edged sword, with question marks over how that kind of energy use is sustainable but also guiding investors in the direction of companies attempting to find ways to meet the demand.

“We're going to see massive exponential increases in energy demand and how is that energy going to be created?

“That demand may mean we need innovative solutions, and AI may help us with solutions for low carbon, renewable energy at scale.”

Valuations on the up

Berry expects companies broadly to become more profitable as they use AI streamline functions or improve their supply chains, bumping up valuations.

“I was talking to a large global consulting business, and they were saying the output of the AI-generated due diligence reporting for companies and strategic reporting for companies was as good as the human reporting.

“Now they hadn't launched it to market yet, but they were going to launch it at a 25 to 30% discount to human-generated.

“The quantity is the same, and it would be a whole change in the category they operate in, in terms of a massive competitive advantage.

“So some companies like that will get a step change advantage and it may not be permanent, or it may put the competitors out of business, but there is going to be that sort of disruption.”

No rush to “AI-ify” everything

Pathfinder is taking an “easy-does-it” approach to AI, first nailing down policies and protections which designate AI as a tool to be used and not a replacement for humans in the business.

“When properly harnessed and incorporated to process the data insights, the time saving on research will be huge, portfolio monitoring, scenario testing of portfolios, there'll just be process efficiency and potentially really good idea generation as well that humans can test those ideas for new investments,” says Berry. 

“AI has to lift what we're doing further,” he says, “because ultimately, people come to us for our duality of returns and ethical investment so whatever we do has to improve on those and it's not an exercise in stripping costs out.”