AI Ethics: Why It Matters for Your Future (And How to Get Involved)
Published 4 months ago • 6 min read
VIEW IN BROWSER
Riley Coleman
August 2023
Why you should give a damn about AI ethics
In this issue
Why it matters
How it impacts your life
Why time is running out
G’day!
Lets talk about how we take advantage of AI, while minimising the risks. Imagine waking up tomorrow. An AI has decided you're unfit for your job, ineligible for a loan, and a security threat. It made this decision using unknown algorithms and personal data you never consented to.. Sound far-fetched? It's already happening, and it's just the tip of the iceberg.
AI isn't some far-off future technology. It's here now, and it's making decisions about you every single day. AI is silently shaping your world. And here's the real kicker: once it learns about you, there's no going back.
Chief Decision Scientist, Google Cassie Kozyrkov explains:
“Trying to remove training data once it has been baked into a large language model is like trying to remove sugar once it has been baked into a cake. You’ll have to trash the cake and start over…
Furthermore your data has been used to train a GenAI model, data deletion cannot be (feasibly) guaranteed even if it is your legal right as a user…
No one has yet solved “catastrophic unlearning” which is when your model’s performance degrades exponentially if you start trying to unlearn/delete huge numbers of datapoints.”
Here’s a real example of this in action. In 2014 Amazon began developing an AI recruiting tool to automate the hiring process. By 2015, gender bias in the company's ratings was clear. This prompted a reevaluation of their candidate assessment process.
You see, AI was trained on resumes submitted to Amazon over a 10-year period, most of which came from men. As a result, the AI learned to prefer male candidates and penalise resumes from women. Despite efforts to re-train the AI for gender neutrality, it found workarounds. It rejected applicants with "women's chess club captain" or degrees from two all-women's colleges. There was no guarantee it would not discriminate in other ways, so Amazon ultimately scrapped the system in 2017.
Well FiretrUCK!!! What are the other options?
Danger - High Voltage sign
To ensure AI serves humanity, not the other way around, I believe in ethical, Human-Centred AI. It can augment our abilities, protect our rights, and improve our lives.
Let me be clear, I am not anti-AI, quite the opposite, but I am also pro-human. Our future living alongside AI isn't guaranteed.
That AI feature you are designing, that new AI HR system you are procuring, or that AI system you are training. Every decision you make is a vote for ‘Team Human' or 'Team uncontrollable, unfair AI'. It's up to us to shape it.
Let's break down each ethical principle and what is at stake for you personally:
Human Agency and Oversight
Human oversight is vital for AI systems. It ensures they respect human freedom, stay in check, and align with our values.
This enables you to make informed decisions, recognise fair treatment, and report AI errors.
AI systems earn your trust through robust data protection.
This means your personal information stays secure, safe from leaks and hacks. This peace of mind enables confident AI use and data sharing.
Privacy and Data Governance
Privacy-conscious AI emerges from ethical design. It respects data laws while delivering innovative solutions. This method ensures both compliance and user trust.
This empowers you to shape how AI impacts your life, ensuring your digital footprint aligns with your preferences and privacy concerns.
Societal and Environmental Well-being
AI systems profoundly impact both society and the environment. AI can worsen climate change due to high energy use. It also impacts social interactions, work, and behaviour. So, balancing innovation with social good and environmental care is crucial.
This ensures AI improves our world instead of harming it.
Accountability
AI firms must answer for their creations' choices. Beyond code audits, true accountability means holding companies responsible when algorithms affect real lives.
This mean you have the right to seek meaningful redress for real-world harms from the companies that deploy unfair AI systems
Transparency
True transparency empowers users to know when they are using an AI. They should understand how it makes decisions. With this knowledge, they can make informed choices.
Without transparency, you can't understand AI's decisions or what personal data it uses about you. You lose your human agency and ability to oversee its actions.
Diversity, Non-discrimination, and Fairness
AI systems can amplify societal biases, harming marginalised groups. AI bias comes from two main sources. The data it is trained on and the algorithm designers' worldview. Ensuring fairness requires parity in data and diversity in teams.
However current state data and team diversity are :
There are over 7,000 languages on Earth. The top 12 languages make up 85% of internet data. English accounts for 52% of all internet data.
We lack global stats on women in AI. However, in 2023, women make up only 29% of STEM workers, as per the World Economic Forum. This includes science, technology, engineering, and mathematics.
The figures raise concerns for anyone who isn't a white man in an English-speaking Western country.
Privacy and Data Governance
AI systems rely on vast amounts of data, raising critical privacy concerns. Strong data governance and privacy are key to building user trust and meeting regulations.
This ensures you have the right to choose what personal data is collected and used with AI. Remember these are learning systems that can't unlearn your data, so be mindful of the privacy standards.
These principles aren't just technical guidelines. They're a safeguard for your future, your opportunities, and your children's world. When we fail to implement these principles, the consequences are deeply personal.
We have limited time to install AI ethics.
Here’s why we can’t wait for legislation or court cases to make these decisions for companies. "We're on a very short runway," cautions an industry leader. This isn’t like other technology systems we are used to. These are learning systems. Let me put this into perspective.
It's the start of the preschool year, and one little kid starts exhibiting some anti-social behaviours. Teachers try to tell the little kid not to take toys and to stop pushing the smaller kids in the sandbox. But, the kid's behaviour doesn't change much.
School yard bully
Because their whole life, they have lived in a household with different, looser rules. They could do what they wanted, go to bed when they wanted, and say what they wanted. And they go back to that house every night. So, it's no surprise that this kid's behaviour doesn't change. They are coded with a different set of ethical rules, and it's been reinforced.
The AI systems built right now are learning and evolving at an unprecedented rate. An expert says, "The decisions made now will shape our future for generations."
But here's the thing: it's not too late. We all have a choice. We can create AI that empowers us, respects privacy, ensures fairness, and aids society and the planet.
However, this won't happen automatically.
The time for passive observation is over. Every line of code, every design decision, every product feature in AI development adds up to shape the systems that will influence our lives.
What can you do to be part of Team Human?
1. Stay Informed: Understand how AI is used in products and services you use daily. Knowledge is power.
2. Speak Up: Don’t be afraid to raise your concerns within your design, development or HR teams. Your voice matters.
3. Design Transparency into AI Features: Support initiatives that promote transparent AI. Your users have the right to know how AI is making decisions about them.
4. Advocate for Diversity: Push for diverse representation in AI teams and training data. AI should understand and serve ALL of humanity, not just a select few.
5. Support Ethical AI Policies: Engage with internal teams and advocate for ethical governance. Also reach out to policy makers, ethicists or experts when needed. Your future is being written in legislative halls as well as tech campuses.
6. Choose Human-Centred AI : Support companies that prioritise ethical, human-centred AI. Your choices as a consumer have power.
7. Spread the Word: Share what you’ve learned with friends, family, and colleagues. Building Team Human starts with awareness. ('cough' - share this article on LinkedIn )
Remember this choice is only available if we act now. It’s time to pick up the paddle and start steering toward an ethical, human-centred AI future.
Join Team Human today. Your life, your opportunities, and your children’s future depend on it. Together, we can ensure that in the age of AI, humanity doesn’t just survive - it thrives.
