Mission

Search

What Does an Ethical Approach to A.I. Look Like?

Play episode

 

The spirit of innovation is mostly positive. After all, innovation  leaps humanity forward. The wheel enabled transportation, space exploration led to the internet, the smartphone has connected the world. These are just some examples.  And the innovation wheel keeps turning. These days there is no doubt that. For our brief history of breakthrough innovation, we’ve rarely had to discuss or think about the ethics behind those efforts like we do with the next breakthrough; Artificial Intelligence and Machine Learning.

A.I and Machine learning  are widely considered to be the tools that will leap humanity forward into the future, but there’s a catch…… who decides the ethics of AI and ML? Who tells the computer how it should think? What should the computer value? What is ethical? And at what point have we gone too far?

“This technology is not fully developed, It’s not an end state. So we don’t know all of the consequences of using this technology.”

When we think about innovation, we think about all the good that will come from it, and rarely think about the consequences that innovation leaves in its wake. Someone is thinking about that, though, and that someone is  Beena Ammanath. As the Executive Director at Deloitte’s Global A.I. Institute., It’s her job to go through all the “what ifs” of A.I.  On this episode of IT Visionaries, Beena explains how she weighs all the outcomes, and she discusses the three paths companies are currently pursuing ethically A.I. She also talks about why trustworthy A.I. and machine learning will be the secret to all successful technology breakthroughs. Enjoy!

Main Takeaways

  • Can I Trust You?: Trust must be at the center of all your A.I. models, which means you need a clear understanding of the data sets you are using and if the data you are using to build your algorithms is reliable. When using third party data, make sure you have a clear understanding of how that data was collected, what the subject matter was and if it truly fits into your modeling. When you can’t trust your data  sources, you end up with biases in your algorithms.
  • The Secret Sauce: A.I. and machine learning continue to be the two big underlying pieces of technology that businesses are using today because of their ability to consistently digest data and learn on the fly. Traditional software used to rely on updates that may arrive every six months, now machines can continually be evaluated and taught new techniques at a moment’s notice.
  • Driving Adoption: Getting people to use your product is always goal number one, but with A.I., and really any new form of technology, consumer adoption is key because in order for A.I. and machine learning to be successful, it’s reliant on the continuous feedback loops it gains from its users. When you’re designing UX, you need to think about how you are going to drive adoption upfront, and not just about how the technology is going to be deployed.

For a more in-depth look at this episode, check out the article below.


Article 

The spirit of innovation is mostly positive. After all, innovation is constantly thought of as what drives humanity forward. Without it, the wheel never would have been invented, we never would have sent an astronaut to the moon, and the internet wouldn’t be able to connect us all from the comforts of our homes. 

And the innovation wheel keeps turning. These days there is no doubt that A.I. and machine learning are having their innovation moment, but while both are expected to deliver tremendous value in the foreseeable future, what are the rules of the road for A.I.? What’s ethical? And at what point have we gone too far?

“This technology is not fully developed, It’s not an end state. So we don’t know all of the consequences of using this technology.”

When we think about innovation, we think about all the good that will come from it, and rarely think about the consequences that innovation leaves in its wake. Someone is thinking about that, though, and that someone is Beena Ammanath. As the Executive Director at Global Deloitte A.I. Institute and A.I. Ethics at Deloitte, it’s her job to go through all the “what ifs” of A.I. On this episode of IT Visionaries, Beena explains how she weighs all the outcomes, and she discusses the three paths companies are currently pursuing when it comes to how they approach A.I. from an ethics standpoint. She also talks about why trustworthy A.I. and machine learning will be the secret sauce to successful technology strategies moving forward, and how technology can drive adoption with a simple UX

Ammanath joined Deloitte full-time in 2019 after starting Humans for A.I., an organization founded to increase the number of women and minorities in A.I.. Now with Deloitte, she’s helping to build ethical models when it comes to how A.I. and machine learning algorithms are developed. 

Ammanath says there are currently three ways to look at A.I. and machine learning; the first is research, the second is taking core technology and applying it to real world industries, and the third is the consequences and risks associated with A.I.

“There are a lot of risks associated with A.I.,” Ammanath said. “I think a lot about that third stream, because I do think we need to be talking about risks and ethics and thinking about how do you put these guard rails in place so that we can innovate faster, and so that we can keep getting value. But at the same time, we are mitigating the negative things that would come out of this technology.” 

When it comes to industrial use cases, today A.I. and machine learning is being deployed in a bevy of different ways, including measuring the effectiveness of continuous running machines, collecting data to ensure the effectiveness of jet engines, and also delivering personalized detail reports on warehouses. But to do all this, Ammanath said that A.I. is not the only piece of technology being used.

“The big thing that we’re using with A.I. today is machine learning,” she said. “[Machine learning] is consistently evaluating and improving. Think of it as a child the first time it’s out there interacting with you. It is still trying to understand datasets, but it’s getting the feedback and learning and improving over time.”

Continuous feedback loops are important, especially when it comes to predictive modeling techniques. With machine learning, Amannath said they are able to continuously measure thousands of data points while also improving their efficiency over time. Something that could not be done with traditional software

“Machine learning continues to learn and evolve,” Ammanath said. “Unlike traditional software 20 years ago, if you built out a program that performed consistently and did the same formula every time, that formula is not evolving based on the data that it is being fed. So it is no longer about a fixed set of core giving you a fixed answer every single time.”

To hear more about machine learning, how Deloitte is utilizing the technology, and how to drive user adoption, check out the full episode of IT Visionaries!


To hear the entire discussion, tune into IT Visionaries here

Menu

Episode 281