There is much talk in the design community about the use of AI for analyzing and critiquing design solutions.
Basically, clients are asking AI to review your design work — be it architectural, graphic, package, product, experiential or otherwise. Clients are using AI to look at a given design solution and depending on their query, they receive detailed feedback and recommendations which they hope to plug back into the design process in order to improve the outcome of the product and reduce risk of failure.
There are pros and cons to this new step in the design process.
Perhaps the biggest pros are:
Definition — That the AI, in order to do an adequate job at creating a critique and recommendations, requires a well described question. This means that the client needs to dig deep to frame the problem. A step that quite often they skip at the start of a design project, typically leaning on the designer to help fill in the gaps in the design brief, or simply move forward on the assumption that the designer simply "gets it".
Levelling up — A second pro is that the critique and recommendations introduce sophisticated language around which the team can have potentially valuable conversation. It can inspire a level of attention to the designer’s rationale that was often previously dismissed by the client as overkill or self serving mumbo jumbo rationalization that they didn’t have time or patience for.
The Cons
The cons on the other-hand are interesting and in themselves great conversation to have with the client because it reveals the true worth of the designer as a problem solver:
Emotional — The design was intended to evoke human emotion and engage with human interaction, both being objectives that are immeasurable by the machine and not available as a data set that could be analyzed and compared by AI to other design solutions
Incomplete Brief — The solution was designed to solve a very specific problem, for an identified audience, using an agreed upon brief however no brief is as informative as the designer’s deeper knowledge of the real nature of the gritty tactile analog world the solution is intended for, therefore the AI can’t judge the solution based on the actual criteria and considerations
Audience Awareness — The solution was designed for a specific audience which the designer maintains a depth of knowledge of which can never be detailed in published data available to AI, and which is generally deeper than the client understands and might convey to AI.
Blind to Invention — Most often the designer understands the problem in a much more holistic way than the client and in particular, the designer understands how to reframe the problem so that the solution is something that doesn’t yet exist — invention — for an audience that don’t yet know they want or need the solution — creating a category — with the goal of sparking a culture and community that still needs to be built around the product — defining a new market. Therefore the client can’t inform AI well enough nor can the AI understand the invention or the value of uniqueness.
Storytelling — The solution was designed to express a specific message, using years of storytelling experience that’s been honed on successful feedback loops with real people but the AI can’t appreciate the emotional value of storytelling
Context — The solution was designed with a heavy dose of context — creative within the human constraints of budget, tailored to the client's sophistication and their ability to leverage and operate the design solution’s system — meaning what we design for a small local inexperienced client needs to be different than an international client with a large sophisticated team — but the AI critic can’t consider the solution as it relates to the client’s capabilities and budget constraints because these realities and affect on the solution are not understood by the client
Currency — The solution was designed using decades of experience that influence a deep understanding of human emotion and an immeasurable intuitiveness that arrives at a solution that was not meant to be compared with yesterday, or today necessarily, rather the intention is that it should be relevant to tomorrow and many years to come — AI can only reference and summarize design solutions which have already been done, pulled from the whole body of work without a curated or stylish eye to discern what is relevant at the moment and adjusted for what might come tomorrow
Empathy — Much of the designer’s job is to provide solutions based on life experience and empathy; things felt during interactions, emotional understandings of the brand, product or space gained during engaging in and experiencing the solutions — AI cannot access the accumulated life experiences shared by humans; the feelings related between audiences, users and guests gathered during face to face interviews or observations while sipping a coffee at a café or strolling the grocery isles — so solutions are not gauged on the merit of connection with human emotion
Effectiveness — Designers provide solutions that have measured successes and wins but which are never shared publicly such as increased margin, audience growth, deepening social engagement, leaps of trust with audiences resulting in stronger conversions — none of which is available to AI to underpin its assessments and recommendations — which means that all it can compare, while missing the human emotion, the experiential and the business end — is the comparable aesthetic and written text of academics that extol on the merits of dated design and trends caught in yesterday’s time freeze — isolated to the superficial and detached from the street level reality of the forward moving design industry
So there you have it. Next time, or the first time, you find yourself within an AI driven feedback loop, hopefully you can host a healthy conversation using these points that will re-inforce your value as well as fend off irrational influences on the project so that you and the project can stay on track.
Cheers