AI — Critical Questions Unpacked
How do you think AI regulation frameworks like the EU AI Act might affect the development and deployment of large language models and other AI capability?
In my reading of legislation like the EU AI Act, there is a fundamental desire to protect people from the worst possible uses of artificial intelligence. A big focus of the Act is on banning or restricting the use of AI to profile individuals – this is a response to companies which have developed AI facial recognition, capability which continues to fall into legal grey areas. But, if nothing else, the Act should give some insight into the incredible power of this technology and how we need to manage its development carefully and systematically, even if that means slowing the pace of its deployment down.
Innovation absolutely needs to be balanced with regulation and care for living, breathing humans and their right to a dignified life. These are weighty concepts, but I truly believe we have arrived at a new juncture in technological development which will bring sweeping changes to how we operate – in how we work, yes, but also in our private lives. As we identify the value of this technology, I do think its use and integration will become more normalised, and maybe even fall into the background somewhat – forming just a portion of an organisation’s wider digital transformation activity, for example. But the broad capability of this technology will definitely change the shape and nature of at least some of the workforce – most AI tech leaders acknowledge this – and also create opportunities for exploitation. This needs to be managed very carefully and needs agreement at a global level.
With the growing concerns around AI-generated content, what do you think are the most effective ways for social media platforms to handle content authenticity?
The smartphone app Halide offers a function called Process Zero which disables all the automated processing a phone performs when you take a standard photograph. The result is something rawer and more film-like - a more authentic capturing of light and colour. Halide already had an enthusiastic user base, but this feature got a lot of attention because it puts control squarely back into user hands. The issue it highlights is that we are already incorporating AI and automated processing into our daily activities, and have been for some time - either without us being aware or caring. My take on this is that automation will only continue down this path, so it will be less about platforms managing the AI content vs non-AI content dilemma than it will be up to people themselves to place a higher value on human-made content. It might look something like choosing to buy a handmade rug as opposed to a factory made one - the masses probably won't care either way, but there will be niche audiences which seek out the hand painted or the handwritten or photography that uses natural light and framing to capture its subject. Human touch or the artificial – which will you choose?
How are you integrating AI tools into your workflows, and what measurable improvements have you seen in performance?
The more I incorporate AI tools into my processes, the more I am finding that the start and end points remain the same. It’s the areas in between that have seen the most improvements and efficiencies through AI integration.
The way I start a program of work is still through identifying or responding to a business problem or need. The problem must be defined, with goals, objectives and desired outcomes then identified. I will then use language models to plan project timings, budget, ideate and even begin my research, feeding it my own ideas and asking for feedback or elaboration - using a language model like Claude has drastically accelerated this part of the process, giving me more time to explore concepts which may be unexpected yet deliver successful results. I also use it as a collaboration tool, testing out ideas and concepts, following thoughts or threads I find interesting. Once I have something more concrete and that I am happy with, I will use AI workspace or canvas tools to begin to visualise and map out my ideas. It’s also at this point that I will share this with others for input or feedback. As the process moves into a production stage, I will use AI for validation of documentation to ensure the concepts are robust, easy to understand and align with initial goals.
The speed that AI allows me to ideate and operate at is a quantum leap - by integrating it into my processes in ways that work for me while also having clear guidelines around its use, I am able to be creative while ensuring I am delivering a high value outcome.
—
Tristan Foster is a digital technology specialist experienced in software delivery, digital transformation and AI implementations in banking, NFPs and startups. He is Co-founder and Principal Consultant at Grand West Consulting and consults with Team DBZ.