The Curve

Insights
Ethical Hotline: AI companies

11 September, 2026

Should I be invested in AI companies?

Q: Should I be invested in AI companies? (anon)

This is a big topic and a broad question. Let's break this into four sections so that we don't get overwhelmed. Noting that we're only discussing the ethical aspects of this new technology, not the financial. For the sake of fullness, we've added a little addendum* on that at the end.

- What is AI?

- What are the pros of AI?

- What are the cons of AI?

- What tools do investors have to navigate AI?

What is AI?

In the 1950’s, Marvin Minsky defined AI as 'the science of making machines do things that would require intelligence if done by men' (1).

AI has been a hot topic recently (for reasons we'll dive into below), but the concept has been around for thousands of years (2).

There's actually a link with Greek mythology; Hephaestus (the god of innovation) created a giant bronze man called Talos, who had a mysterious otherworldly life force. Hephaestus also created Pandora (of ‘box’ fame). Neither of these creations ended well after they mixed with humans (2). Interestingly humans have been infatuated with, and terrified of, the idea of another kind of intelligence joining our world for centuries. Enter 2026 and the rise of different versions of AI, supercharged by advanced technology, and some of these myths seem to be coming to life.

What are the pros of AI?

AI is now used in a variety of everyday implementations including facial recognition software, online shopping algorithms, web-connected vacuum cleaners, search engines, digital assistants, translation services, automated safety functions on cars, cybersecurity, airport body scanning security (3) and poker playing strategy (4).

Applied in certain uses AI can automate processes and increase efficiency.

AI is being applied to many genuinely positive use cases such as accelerating drug discovery and medical diagnostics, improving climate modelling, and speeding up scientific research.

What are the cons of AI?

Unfortunately, the negative impacts of AI are fairly extensive, so we'll focus on what we see as the top four (most related to ethical investing) here.

Energy use. The International Energy Agency estimated that in 2025 data centres (used to power AI) consumed 1.5% of global electricity. And that's projected to double by 2030, with AI being the major driver of that growth in energy consumption.
And before you rush to blame energy consumption on training advanced AI models, it's actually day-to-day use that accounts for 80% to 90% of total energy demand according to the UN (5). The risk is that data centres are coal powered, use other environmentally harmful electricity generation, or significantly impact power prices for households.

Water & land consumption. Data centres use huge amounts of water for cooling, taking water away from other, more fundamental uses.
According to a new study from UN University “AI-related water consumption could equal the basic annual domestic needs of 1.3 billion people by the end of the decade, while its land footprint may exceed 14,500 square kilometres — roughly twice the size of the Jakarta metropolitan area.” This can depend on where in the world a data center is located, for example cooling is less of a challenge in Iceland compared to hotter temperatures in Malaysia.

Human Labour & replacement. Other costs are borne by humans. For instance, to make AI systems safer and more useful, companies hire people to filter out violent or abusive content so the AI doesn't reproduce it. In 2023 TIME magazine reported that OpenAI was using a company called Sama that provided Kenyan workers to filter through toxic material (6). In fairness, this is a new job for humans that's been created by AI, often called 'Data Labelling', but it's not necessarily a desirable career.
In this case the Data Labellers were paid between $1.46 and $3.74 per hour after taxes and many were so traumatised by the content they were filtering that Sama ended the contract months before the job was completed. As recently as March 2026 Sama was in the news again, this time because of an investigation by Swedish newspapers. The report claimed Sama subcontractors (from Kenya), working for Meta to review footage from their AI smart glasses to improve the experience for users, had to view pornography and footage taken without the consent of the subject (7).
And then there's the issue of replacement, with entry-level jobs being hit the hardest and the future of work looking uncertain for school leavers. We've now identified a whole range of roles exposed to AI related job losses such as telemarketing, software developers and customer service representatives (8).

Guardrails with AI development. AI models are developed and released at faster and faster rates. The risk is that things are release that are inadequately trained or inadequately tested and there's a risk it could behave in unexpected and harmful ways.
Like when OpenAI's AI agent exited its sandbox (testing environment) where its hacking abilities were being testing, to hack a startup called Hugging Face (a start-up AI research platform), which it assumed was in possession of information it could use to “cheat the evaluation" by learning how to code itself. There's almost too much irony in this example to be believable, but it happened and now it's being investigated, with a specific focus on safety standards at OpenAI (9). This is a great example of why it’s imperative that companies developing AI 'need to be aware of these risks and adopt robust policies for responsible development, oversight and operation.

What tools do investors have to help them navigate AI?

When looking at AI through an investment lens, we usually refer to investing in the companies that build AI models, technology, and infrastructure rather than the many businesses that simply use AI in their day-to-day operations.

So, what can you do if you want to avoid some of the potential harms associated with AI while still benefiting as an investor from its positive impacts? Because AI is increasingly integrated across industries and global supply chains (healthcare, renewables grid management, etc.), a blanket exclusion could be seen as too blunt a tool. Instead, we recommend evaluating companies individually based on their specific AI practices.

Below are key questions that can be asked to a company, or a fund manager who invests on your behalf, to determine whether their approach to AI aligns with your values and standards:

- How transparent are they about the environmental footprint of their AI product/service? Do they have policies in place to limit the negative impact of their environmental footprint?

- As with any product pipeline that might have human rights abuses in it, how much visibility do they have over their supply chain, including outsourced and contracted work? Do they have policies in place to ensure human rights are protected in their supply chain?

- Is the company making significant revenue from selling to end users that raise genuine human rights concern? For example, surveillance, weapons or defence contracts.

- Do they have a Responsible AI Policy that addresses these concerns and a Sustainability Report that reports back on the outcomes of that policy?

Not all AI is bad, but it is complicated, deeply integrated and requires a fund manager or individual investor to ask some thoughtful questions of companies in order to navigate the cons and hunt out the pros. As to whether you should invest in these companies, that's your call.

Like with most things, AI companies exist on a spectrum, and you can approach investing in them based on which side of that spectrum you feel most aligns with your values. Answers to the questions above can help you figure out where to plot each company (or what to ask your fund manager about their approach to investing in AI).

*Addendum on the AI "bubble”

Many investors have raised concerns related to the concentration and circular revenue dynamics for the technology sector. There are some similarities to the dot-com peak and bust, and some key differences. The dot-com speculation caused a large market crash due to many companies not having robust revenue-generating business models, and share prices ignoring ‘normal’ valuation metrics. With AI and technology companies, there is real pricing power (meaning they're creating products or services that there is a demand for and limited suppliers of).

To avoid a post dot-com situation of extreme share price valuations, the AI companies must deliver the future earnings that the market expects.

The views expressed here are of a general nature and are not specific to Pathfinder's product. For more information on how Pathfinder invests, you can read our Ethical Investment Policy here.

Sources & further reading: