What should be the role of AI or Generative AI in our lives? If the AI were another person, how would it relate to us? And how should we deal with it?
Understanding the roles that we want our AI products to play can shape our UI and even the business models around it.
So far, I’ve been able to think of 4 roles:
- The people-pleaser
- The project partner
- The freelancer
- The guide
Let’s expand each one of them and see what they can shape into.
The people-pleaser:
This is the most visible role we see AI products playing today. The relationship between humans and AI is akin to trying to please someone who never says yes. Whatever the human says, the AI is trying to do its best to agree with it, comply with it, and deliver on it. No cross-questioning, challenge, criticism, or even a nudge for the human to put in extra effort. Barring the hard-coded filters and guardrails, the AI does not care about the request’s correctness or the computational cost of satisfying it, or about the effort the user puts in.
This creates the problem of echo chambers. Finding and latching onto data that reinforces an individual’s beliefs becomes even easier. An individual’s ego gets bloated without checks.
The project partner:
The relationship between the human and the AI in this scenario will be more like that of class project partners. The tasks are supposed to be divided beforehand. Both the AI and the human are allowed to build on top of each other’s work. Both put in proportionate effort toward the same goal. This is where AI should be able to reward your efforts or thoughtfulness of requests.
It expects you to pull your weight toward a shared goal you decide with the AI. Agentic workflows have an excellent opportunity to become strong partners for humans rather than merely executors of tasks. A thoughtful project partner like AI can be a valuable asset for individuals, especially for knowledge transfer.
The freelancer:
For those who’ve never hired a freelancer, let me outline the typical process. You find an individual with a technical specialty. You give them a brief about the project or expectations. If they are any good, they do not immediately jump into execution mode; instead, they ask follow-up questions to clarify the project scope. They clearly understand your requirements, define the must-haves and nice-to-haves, and, once the scope of work is fully defined, propose a timeline and a price quotation.
AI can establish consensus on the initial creation step, gather feedback, and continue building on the suggestions. If an AI acts like a freelancer, there to assist us responsibly, we will likely also try to be more sophisticated in our requests. This will be difficult for companies to do because generative outputs are somewhat unpredictable. However, this is the one we should aim for.
The guide:
The current state of AI can also serve as a helpful guide. Instead of directly responding to a user’s request, it can probe deeper, ask for context, and even question the need for the thing the user is trying to create. This would require AI to be trained from a completely different standpoint, where the process is more important than the outcome. This kind of AI may not have clear financial value on day zero, but it can be instrumental in improving human quality of life.
Most AI today is built to please. The next wave will be built to collaborate, scope, and guide. The idea of anything other than a people-pleaser is uncomfortable to us because we view AI as a non-living tool created by humans. We at Pikoo believe the better answer lies somewhere between freelancer, partner, and guide behaviours—different behaviours for different use cases.
For our AI game creator, we’re heavily leaning on freelancer behaviour. Scope out the problem and get as much clarity on the idea stage as possible to avoid wasted effort in execution. We are experimenting with partner behaviour with our game jam tool, where we decide on a mutual target of creating a game, and we bring more to the table as the user brings in their ideas. The guide behaviour is being tested in our game design document creator tool, where the user’s initial assumptions will be challenged by real-world information and ideas as they create their game design document step by step.
We’re excited about all these experiments and look forward to how the users respond to them. We’ll keep you posted as things pan out. But we’re certain that the future of AI is NOT just more output. It is actually better, more intentional, and context-aware relationships.
Leave a Reply