ITP Techblog

Business is waking up to the ethical dilemmas posed by Artificial Intelligence and it's time to move the discussion from the tech community into the mainstream. That's the prevailing message from the report by Chartered Accountants Australia and New Zealand (CAANZ), an organisation representing 120,000 public accountants.

The report, entitled Machines can learn but what will we teach them: ethical considerations around artificial intelligence and machine learning provides a summary of the issues that have been brought to forefront in recent times. These include:

• The impact of AI on employment - the OECD predicts around 66 million job globally are at risk of automation.
• The effect new technologies are having on cognitive learning functions as users spend more time on devices and "no one yet fully understands how they are shaping our ongoing human social and physical evolution".
• Fake news, hate speech and the manipulation of data having real-world political consequences. In April 2018, a Pew Research Centre report analysed 1.2 million tweets and discovered that two-thirds came from suspected bot accounts.
• Potential bias by AI designers and programmers as there is a notable lack of diversity in tech - "women and ethnic minorities, for example, are greatly under-represented in the coding arena; will this result in a gender or ethnic bias in how AI reaches its recommendations?"
• Absence of transparency around the algorithm in use - the idea that it becomes a "black box" so that no-one knows exactly how it is arriving at its decisions

As a profession, accountants are well-placed to understand the effects of AI, having had to pivot to become trusted business advisors as accounting software is "getting more intelligent, performing automation as well as analysis previously done by humans". Therefore it's interesting to see how its industry body CAANZ is tackling the issue of AI in a report that grapples with the idea of creating an ethical framework around how AI is applied.

"If we fail to build the ethical dimension into each stage of our AI journey, an alternative route that perpetuates the current polarisation of wealth and resources within and between societies seems inevitable. The academic world has put in place an ethics regime around research with humans which may be the appropriate starting point in considering the type of ethical framework necessary to guide ongoing AI developments."

The report notes that despite the widespread use of drones, fingerprint technology, facial recognition and driverless cars, there has been no commonly agreed accountability framework, although governments and regulators are looking for ways to deal with the increasing use of AI. "Most national governments understand that over-regulating AI would be a simplistic reaction rather than a strategic response to AI advances and that if they over-regulate, AI projects and further investments in AI will simply be moved to more relaxed regulatory regimes."

Increasingly businesses and organisations are expected to be mindful of their socially responsibility when designing products and protecting customer data, and this may require them to create AI ethics code. Individuals too will need to understand the impact of AI, as the report notes: "The anonymity provided by the internet is not ethics free and brings with it the same ethical personal accountability to knowingly "do no harm" to others."

You can read the full report here.