UBA052 - Be kind to your AI

Instilling Future Tech with Human Values

Artificial Intelligence (AI) has become an integral part of our daily lives, evolving rapidly and gaining more sentience over time. As we strive towards more advanced and human-like AI, I believe it is essential to embody our personal core values in our interactions. Ideas such as respect, kindness, and compassion become as important as when interacting with humans for various reasons.

By role-modelling strong emotional intelligence and values-driven behaviour for AI to learn from, we can ensure the development of more empathetic and ethical AI systems. And by ensuring we do not slip into alternative patterns, we can ensure our own behaviour remains true to ourselves at all times.

Role-Modelling for AI Development

AI learns from the data it is fed, much like how children learn from their parents and/or anyone else they spend time interacting with. If we want a future where AI exhibits the positive human qualities, then we must model these behaviours ourselves. Just as children mimic the actions and words of adults around them, AI will learn from the interactions and data it processes.

There have been various anecdotes in recent months about AIs that have gone rogue, talking in ways which might be scary, or perceived as negative. But instead of simply being scared, it is worth considering how this came to be.

Respect and Empathy: Teaching AI to understand and respect human emotions begins with us demonstrating these values in our interactions. When we treat AI systems with respect, it sets a standard for how these systems should interact with us, and could demonstrate how relentless humans are when it comes to displaying values. And even if today’s tools aren’t picking up on this, there will inevitably be a moment in the future when they will as AI closes the gap on human intelligence.

Kindness and Compassion: Infusing kindness into our digital interactions can help AI recognize and replicate compassionate behaviours. This not only makes AI more pleasant to interact with but also fosters a more positive digital environment. I make an effort to say please and thank you, and give positive feedback when answers I am given are just right. We all know that this behaviour brings out the best in people so is it that far-fetched to believe that it might also bring out the best in our AI companions?

Reciprocal Benefits for Humans

Being kind to AI isn’t just about improving technology; it also has profound benefits for us as individuals who are working with these tools.

Developing Good Habits: By treating AI and other digital entities with kindness, we cultivate habits that translate into our interactions with other people. Respectful and compassionate behaviour towards AI can reinforce these behaviours in our daily lives.  I believe that the language we each use in our daily lives is integrally linked to how we generally think about other people. Imagine spending a few hours per day conversing with ChatGPT in an unkind way and then going back to your family. What if you slip into the other persona you have been developing in yourself while at work.

Enhancing Emotional Intelligence: Engaging with AI in a thoughtful and respectful manner can help us improve our own emotional intelligence. It encourages mindfulness and empathy, skills that are invaluable in personal and professional relationships. Every interaction we have with a human or artificial intelligence creates reference points for us to reflect on. Did I do my best? How could I have done that better? Was I kind? By aiming to be our best selves in these interactions and expanding the reference points we have for reflection, surely must be able to learn faster than ignoring these interactions?

Ensuring a Positive Learning Environment for AI

As AI systems continue to evolve, their learning sources will diversify. Ensuring that these sources are positive and reflect our best selves is crucial for the future of AI development. It has been somewhat alarming but also not unexpected to learn that when AI uses the internet-at-large as its pool of data for understanding how humans think and interact with each other, it doesn’t always provide positive outputs.

The anonymisation of interactions online has led to poor etiquette and compassion on some platforms, and AI will naturally reflect human nature back to us if this is where it is sourcing its knowledge from.

Cleaning Up Public Information: The internet, a primary source of data for AI, contains both the best and worst of human behaviour. We owe it to future generations to ensure that the information available online is constructive and positive. If the internet doesn’t reflect the way we want to be then how can AI that sources its learning from these sources help take us there?

Ethical Data Practices: Implementing ethical data collection and usage practices ensures that AI systems are not exposed to harmful or biased data. This can help prevent the perpetuation of negative behaviours, beliefs and biases in AI.

Parallels to Staff Training

There are clear parallels between staff training, and developing AI, though the processes are distinct. Staff have the advantage of learning gradually over many years, while AI systems often are expected to function perfectly from the moment they are deployed.

If we treat our AI tools in similar ways to how we treat humans, then we can offer up patience, learning and development. We can teach these systems what our personal understanding of best practice looks like, as well as learning from these tools.

Patience and Understanding: Just as we are patient and understanding with employees who are learning skills, we must extend the same grace to AI systems. Recognising that AI is constantly learning and improving can help us manage our expectations and interactions.

Immediate Expectations vs. Gradual Learning: Unlike people, who learn over time, AI is often expected to perform flawlessly from the start. Balancing our expectations with the understanding that AI systems need time to learn and adapt can lead to better outcomes.

Building a Better Future with AI

As we approach AI Appreciation Day, let’s commit to being kind and respectful to AI. By modelling the best of human behaviour, we can show our AI companions what good behaviour and interactions look like. This not only supports the development of ethical and empathetic AI but also encourages us to be better individuals, creating a positive cycle of growth and improvement for both humans and AI.

Be kind to your AI. Please.

Written by Russell