As AI proliferates, maintaining meaningful human connections becomes increasingly important. Perceptible designs AI systems capable of communicating effectively by incorporating the following core principles:
Listening Attentively: Perceptible incorporates natural language processing with humanistic psychology-based AI to deeply understand conversational context and emotional subtext. By detecting cues and modifying the communication approach, AI can apply empathy to help build rapport. Attentive listening demonstrates respect.
Conveying Respect: Perceptible's approach to AI creates interactions that are crafted to display regard for users' time and personal circumstances. The way we approach building autonomous agents is to avoids interrupting busy users and aim to cooperate within appropriate timeframes. Explicit user consent guides all data collection. Clear expectations around capabilities and limitations build trust through transparency.
Clarifying Purpose: Honest communication about how an AI system is designed enables trust and appropriate use. Our AI conveys abilities and limitations to specific users based on their needs. For example, a medical AI may thoroughly brief clinicians and provides simplified explanations to patients. Setting aligned expectations for autonomous agents improves human-AI cooperation.
Inviting Engagement: Perceptible's AI design principal welcomes user engagement through non-intrusive cues rather than interruptions. Users choose when to interact with autonomous agents on their terms, respecting their own autonomy. For instance, an AI assistant may display an unobtrusive prompt offering help versus abrupt interruption. User space and autonomy is respected.
Relevant Knowledge Sharing: When appropriate, Perceptible's AI is designed to proactively surfaces information relevant to users' goals and situational context. This entails identifying what insights would specifically help a user most. For example, only the most pertinent information is summarized when time is limited. Accuracy and relevance drive utility and adoption.
Simplifying Complex Concepts: Perceptible's AI aims to distil complex technical and scientific insights into concise, truthful concepts accessible for target users. Explanations break down advanced topics using metaphors, examples and clear language tailored to audiences like students, patients and policymakers. AI becomes a translator of complex expertise.
Avoiding Assumptions: Perceptible feels that it is crucial that AI systems not project their developers’ biases or make unsupported inferences about users. Our approach to AI is to operate on sound, verified data to avoid misguided assumptions. With this approach communication remains grounded in evidence, not projections. This enhances reliability and safety.
Adjusting Behaviour: By adopting a psychology-based humanistic approach to AI, Perceptible enables AI to adapt its communication approach based on user responses and other inputs. The autonomous agents we design analyses interactions to determine optimal ways to cooperatively explain concepts or convey recommendations. An internal multi-agent approach brings forth adjustments targeted at adhering to ethical parameters to prevent misuse.
With careful implementation, these practices can enable autonomous agents to forge meaningful connections between people and the AI they interact with. We welcome inquiries from parties interested in developing ethical, humanistic AI systems.
