For the Landmark Planning event at the Marquis of Wellington the three speakers James Blood (SGP), Oliver Hatton (Edge) and David Langston (RCS Technologies) were introduced by me, who also managed to throw in a few usual bon mots (we hope).
Each speaker took a different slant on the introduction and use of Artificial Intelligence in their particular sector from their perspective.
What was really interesting was when you get into the detail of what each particular company does in its sector of the development industry the vast range of potential applications that exist. And then, of course, that can be partly replicated and partly lead to new applications in other companies, in other sectors of the industry.
What we were trying to get to is to identify more general truths that have a broad applicability or resonance across the whole industry.
The plenary discussion session was frankly limited. I am not sure if that was just because the stepped change that is coming is such that many people did not feel confident asking the ‘right’ questions. And this is one of the more obvious lessons from the night. To get the most benefit out of AI you have to know the ‘right’ questions to ask and phrase it, so the questions are coming from the right angle to benefit the human customer.
And that leads on to what for me was one of the most interesting points, which came out of the informal chats afterwards with an end user of the work of those presenting their experiences.
“In your list of speakers you didn’t have the customer perspective.” That was a lightbulb moment for me. We are all getting very wrapped up in the technology and our ‘consultancy and what we assume are it's benefits for the quality and speed of our work. We assume it will automatically benefit the customer, but how much and will it? Has anyone asked the customer what they want out of it?
All the presenters very emphasised that it is a tool and you need the human to interrogate and review it before any issuing of information. But you can easily see the temptation to push a project out under time pressure without properly auditing its output, as it looks ‘good’ on initial examination.
And, equally, it needs very careful auditing; both because the questions and the questioner can significantly influence the AI output. There can be a hallucination effect where AI tells the questioner what it thinks it wants to hear. Here at Landmark Planning, we have tried essentially questioning from different input sources. Although not radically different they were different. And when, in our case, we have looked at the subtleties of the material planning balance it does not seem at present to be able to effectively weigh up the relative importance of different issues. This may, of course, improve, as AI learns from our previous work.
And this leads on to a question asked at the very beginning of our plenary session. Whose Intellectual Property (IP) is any output? How to protect your own work, with internal AI structures, which is what companies such as those presenting have set up is only a partial solution. Once it is out there to be accessed in any way by AI, that IP appears to probably be completely lost.
Another thorny question and where it is just too early to make a definite judgement is will it lose or create jobs overall? Most, but not all technological revolutions creates jobs, but is this a hand weavers revolution which decimated whole communities (!800 c. 240,000 hand weavers; 1860 c. 10,000 in UK). Certainly, it is more likely to remove lower skilled jobs, but then do how do young people learn in the first place?
So many questions in search of answers.