At Ethical Pixels®, we’re always excited about emerging technologies, but our mission to promote ethical uses of technology means we also have to think extra carefully about the implications that come with these technologies, and how comfortable we are using them. That’s why this page details how we use Generative ‘Artificial Intelligence’ (AI).
TL;DR Summary
Ethical Pixels® is a human-first consultancy. The work you get from us comes from people and not AI tools.
We actively choose tools without built-in AI components, or (where this isn’t practical) disable and avoid AI components in the software and tools we use.
We might still use open-source machine learning tools for internal processes where we can verify the training data has been sourced with consent, or where the tool is simply more efficient at certain types of processing or pattern matching.
Since ChatGPT and other Large Language Models (LLMs) exploded into the public consciousness, we’ve noticed a rush to implement AI technologies to do anything and everything imaginable. Wether this is born out of an over-zealous desire to embrace new technologies without doing the hard work of considering the implications, a fear of being left behind, a cynical desire to increase the perceived value of a product or service (or some combination of all) is open to debate.
Although we’ve dabbled internally in the use of Large Language Models (LLMs), we’ve come to the reluctant conclusion that AI cannot be truly ethical as a technology at this time for a host of reasons:
- Sources and consent for training data might be dubious, presenting issues with plagiarism and copyright
- What you enter into a model can be used for training the model further
- The ecological and energy impacts
- Potential for disinformation and use to perpetrate increasingly convincing scams
- Human impacts, such as job displacement and cognitive outsourcing
- The potential for harm and inappropriate guardrails
- Impact on the price of technology and subsequent reduction in democratisation of hardware, software and information
- Potential existential risk to humanity
We’ve explored all of these issues in more detail if you’re interested, but for the purposes of this policy, we’ve focused on how we approach AI.
We’re experts in Digital and Web – for us to say that we won’t embrace a watershed technology might seem insane from an outside perspective. And yet pursuing a human-first model has been how we’ve delivered quality, bespoke and stable experiences for the last 7 years.
For us, the choice has been to either reject AI and keep our principles intact, or embrace AI and go back on some of our ethical promises. We can’t currently reconcile both.
The steps we take
- We no longer use commercial Large Language Models or Generative AI tools
- When choosing our technology stack and tooling, we actively choose tools without built-in AI components.
- Where this isn’t practical, we disable and avoid AI components in the software and tools we do use.
- We train our staff on the impacts and implications of AI usage
Exceptions
- We might still use open-source machine learning tools for internal processes where we can verify the training data has been sourced with consent, or where the tool is simply more efficient at certain types of processing or pattern matching. This is useful for processing data rather than generating data.
- Any machine learning or AI tools we do choose to use should be powered by renewable energy.
Understanding “Artificial Intelligence”
The final thing we always remind our clients is that “Artificial Intelligence” is (at least at the time of writing) neither of those things. The words, artwork and video that is used to train these models had to be created by humans originally, so isn’t really artificial. The outputs derived from this might be quickly generated, sound convincing and seem impressive to us, but it’s really just pattern matching and machine learning – not intelligence.
The term has become so misused and confusing that the industry had to come up with an alternative term (Artificial General Intelligence, or AGI) to refer to the search for a genuine machine intelligence.
If you have any questions about our use of generative AI tools, feel free to contact us.