What is it about?
In the fast-paced world of artificial intelligence, we've seen it become an integral part of our daily lives, revolutionizing everything from managing emails to tailoring music and movie suggestions to our preferences. It's a technological marvel, no doubt, but there's a darker side to this seemingly utopian picture; biased or incomplete datasets and a lack of transparency in algorithms are the culprits. In essence, in AI, there is an ethical challenge that defines our century. Developing ethical AI requires more than just a human-centric approach. Despite the ongoing discussions, this approach alone falls short of ensuring ethical algorithms. So, why the controversy? This is where our story steps in. This paper suggests a holistic solution, advocating not only for a human-centred approach but also an "ecological" one. What does that mean? Think of it as considering the context of use—the here and now. Drawing inspiration from Bronfenbrenner's ecological approach to child development, the paper draws a parallel between a child's self-determination through learning and an AI algorithm's enrichment through learning. By placing the child in the algorithm's shoes and experiencing interactions in its surroundings, the proposed solution aims to address the ethical challenges of AI. Concrete evidence of this ecological model's viability comes from a use case in the European research project FRACTAL. In developing AI-based algorithms for advertising and customer assistance in shopping centres, the focus was on gender recognition algorithms known for perpetuating biases against women. While this paper doesn't claim to have all the answers, it takes significant strides in the right direction. It presents a conceptual framework for training and retraining AI algorithms, laying the foundation for a theoretical and practical rethinking of AI algorithm design. It's about combining the human-centred with the ecological approach, moving us closer to resolving the ethical dilemmas posed by AI.
Featured Image
Photo by Markus Winkler on Unsplash
Why is it important?
In the world of AI research, a colossal ethical challenge is shaping the narrative of our era. The urgency of the matter is emphasized by a comprehensive incident database, which meticulously logs over a thousand AI-related blunders (check it out at incidentdatabase.ai). A paradigm shift is proposed to tackle this ethical quandary head-on: enter the ecological model. Considering the real-time context of AI usage, this approach proves to be a game-changer. It's not just about preventing discrimination and bias; it's a holistic strategy to meet end-users needs, enhance user-friendliness, comply with ethical regulations, optimize performance, and more. What's revolutionary here is the fusion of the human-centric and ecological approaches in the design model. This isn't claiming to be a silver bullet, but it's making significant strides in the right direction. The paper unfolds a conceptual framework for the training and retraining of AI algorithms, setting the stage for a complete overhaul of how we think about AI algorithm design. It's all about merging the human-centred with the ecological approach, pushing us one step closer to untangling the ethical knots that AI presents.
Perspectives
With a quarter-century steeped in Human-Computer Interaction, I wholeheartedly assert that a seasoned researcher holds a pivotal role in tackling the fresh ethical conundrums of AI. Having been hands-on in the actuation of Human-Centered methodologies in real-world scenarios, there's a unique perspective here. AI, with its distinct characteristics, stands apart from traditional systems. The AI communities are now doubling down on ethics, pouring in significant investments to pinpoint human-centred principles that should steer the deployment of AI systems. We're standing at a historic crossroads, witnessing the inception of a collaborative journey between the realms of Human-Computer Interaction and AI. In my perspective, donning this responsibility isn't just an ethical duty; it's a distinct honour. It's about melding the wisdom from two domains to navigate the uncharted waters of AI ethics
Tania Di Mascio
Universita degli Studi dell'Aquila
Read the Original
This page is a summary of: How to Make an Artificial Intelligence Algorithm “Ecological”? Insights from a holistic perspective, September 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3605390.3605398.
You can read the full text:
Contributors
The following have contributed to this page
