Articles/Ethical AI: In the pursuit of AI for good

Ethical AI: In the pursuit of AI for good

The ethics of AI can be polarising. Kath Blackham believes that AI can truly create better lives for people but that humans have a responsibility to make sure that systems are inclusive.

As a leader in conversational AI, I speak a lot about artificial intelligence (AI). And there is one question someone in the audience always asks me. Every. Single. Time. It’s the elephant in the room. It’s a question about ethics and AI. The answer? It isn’t about a machine’s ethics. It’s about ours.

Just as there are cat people and dog people, when people think of AI, they usually fall into one of two camps. That AI is evil and the machines are going to take our jobs and take over the world camp. And the AI for good camp, who are excited about greater creativity, productivity and advancing society. 

The pursuit of AI for good

I’d very much like to think that my business falls into the ‘tech for good’ camp. I have made it my life’s work to create better lives for people. AI can help us do that. 

Here are just some examples:

  • AI can do that for the one in six Australians who live with a disability.
  • AI can do that for the one in five Australians who can’t access or understand information served up to them in English, because they speak a language other than English.
  • AI can do that for people who are excluded due to conscious or unconscious bias because of their age, gender, ethnicity, sexuality, demographic profile or the postcode in which they live.

I not only believe it is all possible, I know it. Together with forward thinking client partners, we are already building inclusive conversational AI-based systems for not-for-profits, government and corporates to reach diverse audiences in ways that best suit the end user’s needs and abilities, not the organisations’.

By building voice-based AI systems, people who don’t speak English can ask for and receive information in their own language. Systems that let people with low vision, low literacy, or poor dexterity (including the elderly) access information without reading or typing. A bit of a shameless plug here, we’ve even just built one of the world’s first navigation-less websites so people can ask for what they want by speaking instead of typing in their own words. 

AI is not perfect

AI is evolving technology, and it’s not perfect. It can lead to discrimination and exclusion when the system isn’t built to reflect the diversity of humanity. AI systems have famously failed to correctly classify some of the world’s most famous faces. One was 76% confident Oprah Winfrey was a male. Another was 80% confident Michelle Obama was a young man and 93% confident she was wearing a male hairpiece.

Like kids, AI learns by what we give, feed, teach and show it. Give it a diet of poor data, and you can end up with in-built bias. That’s what computer scientist and founder of the Algorithmic Justice League, Joy Buolamwini, discovered when facial analysis software couldn’t detect her dark-skinned face until she put on a white mask. Writing about her experience in Time, Joy asked, “How much does a person have to change themselves to function within technological systems that increasingly govern our lives?”

You don’t have to change, Joy. The systems do. And that comes back to how we as humans design them. 

The human driving the machine

At the moment, the misconception is that the machine is more powerful than it is. But the reality is, it’s the human driving the machine. The AI is learning, but it’s not making its own decisions. We are in charge. And that means we have a responsibility to make sure the systems are inclusive, representative and don’t discriminate because of poor data or flawed algorithms.

I talk a lot about AI, but I also talk a lot about my life’s work. For me, that is tech for good. I want to show the world how this technology can be used not for evil and discrimination. Not to collect and sell people’s data. Not to encourage people to do things that they don’t want to do – but how it can really be used for good. How the technology can be used to reach all parts of society. The culturally and linguistically diverse, those living with disability and those from demographics that have been largely ignored by mainstream media and marketing and communications, whether that be age, gender, sexuality or postcode. 

As a society, we should all be thinking about how we can use these technologies for good to create better lives for all – not just some. 

The negative always seems to grab more headlines (‘the robots are taking over the world!’) but the positive side to all of this is you can use AI technology to reach the unreachable, to be inclusive and make sure the message gets through to people so that action and change can happen. 

It already is.