There are three standard approaches to data and AI ethical risk mitigation, none of which bear fruit:The academic approach: academics are great at rigorous and systematic inquiry. But they ask questions like “Should we do this?” Businesses ask, “Given that we are going to do this, how can we do it without making ourselves vulnerable to ethical risks?”. The mismatch in language makes it a bad fit to give good answers

What not to do (1/3)

Companies are investing in answers to once esoteric ethical questions because failing to operationalize data and AI ethics is a threat to the bottom line. Despite the costs of getting it wrong, most companies grapple with data and AI ethics through ad-hoc discussions on a per-product basis. When companies have attempted to tackle the issue at scale, they’ve tended to implement strict, imprecise, and overly broad policies that lead to false positives in risk identification.There is a need for risk mitigation plans for AI based solutions that can actually be used.

Ethics in AI are a reality 

In some cases AI solutions go wrong, introducing biases. For example:<ul><li><a href="https://www.latimes.com/business/technology/la-fi-tn-city-attorney-weather-app-20190104-story.html">Los Angeles is suing IBM </a>for allegedly misappropriating data it collected with its ubiquitous weather app.</li><li><a href="https://www.wsj.com/articles/new-york-regulator-probes-unitedhealth-algorithm-for-racial-bias-11572087601">Optum is being investigated </a>by regulators for creating an algorithm that allegedly recommended that doctors and nurses pay more attention to white patients than to sicker black patients.</li><li><a href="https://www.washingtonpost.com/business/2019/11/11/apple-card-algorithm-sparks-gender-bias-allegations-against-goldman-sachs/">Goldman Sachs is being investigated </a>by regulators for using an AI algorithm that allegedly discriminated against women by granting larger credit limits to men than women on their Apple cards. </li></ul>

Sometimes AI backfires

Th high-level AI ethics principles . Google and Microsoft, for instance, trumpeted their principles years ago. The difficulty comes in operationalizing those principles. What, exactly, does it mean to be for “fairness?” What are engineers to do when confronted with the dozens of definitions and accompanying metrics for fairness in the computer science literature? Which metric is the right one in any given case, and who makes that judgment? For most companies — including those tech companies who are actively trying to solve the problem — there are no clear answers to these questions.

What not to do (3/3)

Reid Blackman, the author of this hbr article recommends that organizations take a few steps to ensure an ethical approach:<ol><li>Identify existing infrastructure that a data and AI ethics program can leverage . </li><li>Create a data and AI ethical risk framework that is tailored to your industry </li><li>Change how you think about ethics by taking cues from the successes in health care .</li><li>Optimize guidance and tools for product managers .</li><li>Build organizational awareness.</li><li>Formally and informally incentivize employees to play a role in identifying AI ethical risks.</li><li>Monitor impacts and engage stakeholders.</li></ol>

How to operationalize data and AI ethics

The “on-the-ground” approach . Businesses ask pertinent questions through enthusiastic engineers, data scientists, and product managers. What they lack, however, is the kind of training that academics receive. As a result, they do not have the skill, knowledge, and experience to answer ethical questions systematically, exhaustively, and efficiently. They also lack institutional support.

What not to do (3/3)

Similar ideas to What not to do (3/3)

Building ethical AI