Register now for KMWorld Connect and join us virtually November 16-19!

Coming soon to your newsfeed —Ethics and AI

Ethical challenges

Attention is now being paid to the pervasive role of bias in our technology. We are seeing more and more examples come to light of the problems of bias in the training process required for machine learning programs: Humans are making the decisions about what data to use to make the machine tick, and they have a way of missing ethical implications. Humans are also making decisions about what kinds of algorithms will produce the best results across the data available to the programs, but are “the best results” primarily those the developers are looking for? What about ethical challenges in the business models of our organizations? We see the example of the EU moving to address the unrestricted use of individuals’ data by online businesses seeking to monetize our attention and beginning to extract more than slap-on-the-wrist fines on companies who continue to ignore GDPR regulations (Alphabet).

It almost goes without saying at this point that our social media environment is rife with what all will agree are unethical practices, from Cambridge Analytica to election tampering by outside interests to clearly (or not so clearly) twisted tweet storms.

So, as the hype cycle turns toward ethics this year, there are ways that we can resist the push toward one-dimensional headlines. Fortunately, there are some fine online resources that we can leverage to make ourselves smarter. The work by Michael Sandel at Harvard is very accessible at http://justiceharvard.org. There are also multiple TED Talks that address ethical issues in AI—google “ted talk ai ethics.” And there is the material being created and shared online by the MIT and Harvard researchers—check MIT’s Media Lab Ethics Initiative and the Dalai Lama Center for Ethics, as well as publications from Harvard’s Berkman Klein Center’s Ethics and Governance of AI projects. You will find many more.

Taking some time to catch up on the ethical challenges and gotchas in AI will be time well invested this year, as we are sure to be hearing a lot on this topic throughout 2019, and (hopefully) well beyond.

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues