Artificial intelligence has made many breakthroughs in the last decade, including beating champion players at Jeopardy!, learning to identify cats, seeing better than humans, and driving cars autonomously.

A new study (PDF) by researchers from the Commonwealth Scientific and Industrial Research Organisation’s (CSIRO) Data61, along with the Australian National University and researchers from Germany, has determined that AI can influence human decision-making.

The research involved having humans play three games against a computer.

In the first two experiments, people were asked to click on red or blue-colored boxes with fake currency as a prize. In the third, participants were given the role of an investor and asked to make investment decisions, with the AI acted as the trustee.

As the experiments continued, the AI learned the patterns of each participant and then was programmed to influence their decisions and guide their choices to help them earn more money.

Artificial intelligence, decision-making, and good versus evil

Of course, this raises the usual questions about AI ethics and how we continue to leverage technology for good. This study shows how AI can influence human decision-making and exploit the vulnerabilities of an individual’s habits and patterns.

“The world produced an immense amount of data over the past months,” Elena Poughia, CEO and founder at Dataconomy, said at Data Natives Unlimited in 2020. “It is our responsibility to handle this data with care – staying both private and transparent, sharing our data while protecting it, and always keeping in mind that impact is the new money.”

And this, ultimately, is what will determine whether behavioral change and decision-making AI will be seen as good or evil. It won’t be the AI that is inherently positive or negative, but those in charge of creating the algorithms and the data’s biases.

AI ethics is key

With this in mind, Data61 created an AI ethics institute in 2018. The Gradient Institute is an independent non-profit charged with researching the ethics of AI, as well as developing ethical AI-based systems, focused essentially on creating a “world where all systems behave ethically.”

It explores the ethics of AI through practice, policy advocacy, public awareness, and training, specifically where the ethical development and use of AI is concerned and uses research findings (such as this latest study) to create open source ethical AI tools that can be adopted and adapted by business and government.

Showing that AI can affect human decision-making is another milestone for artificial intelligence. Where we take that is up to us, as the implications of affecting behavior for the wrong reasons are clear and apparent.