digests /

Last Week in AI #46

Ethics, AI for breast cancer screening, and more!

Last Week in AI #46

Image credit: Karen Hao / Technology Review

Mini Briefs

In 2020, let’s stop AI ethics-washing and actually do something

In 2019, there was more talk of AI ethics than ever before. Dozens of organizations produced AI ethics guidelines; companies rushed to establish responsible AI teams and parade them in front of the media. In the wake of a number of ethically concerning facets of AI’s actual applications being brought to light, it now seems that every AI-related conference makes sure to include an ethics-related message as part of its programming.

However, for all the lip service paid to data and AI ethics, many organizations’ AI ethics remain vague and difficult to implement, while few companies can demonstrate tangible changes in how AI products and services are evaluated and approved. For example, Google created a nominal AI ethics board with no actual veto power over questionable projects–after controversy provoked by the inclusion of a few members, the board was dissolved.

Despite these gloomy points, 2019 was the year of the greatest grassroots pushback against harmful AI from community groups, policymakers, and tech employees themselves. While a number of cities banned public use of face recognition, employees of tech giants grew increasingly vocal against their companies’ use of AI for tracking migrants and drone surveillance. AI researchers have also doubled down on mitigating AI bias and reexamined incentives leading to the field’s runaway energy consumption. In 2020, the hope is that industry and academia will sustain this momentum, making concrete changes from the top and the grassroots level that realign AI development.

AI shows promise for breast cancer screening

Screening is used to detect breast cancer early in women who have no obvious signs of the disease. This image-analysis task is challenging because cancer is often hidden or masked in mammograms by overlapping “dense” breast tissue. This difficulty has stimulated efforts to develop AI systems for improving diagnostic performance–Ai systems that outperform expert radiologists in accurately interpreting mammograms from screening programs have already been developed.

Such positive results suggest that AI may someday aid in the early detection of breast cancer. While this sort of research and its potential impacts seem incredibly impactful, there is reason to temper excitement about this and similar AI studies. For one, the performance of AI algorithms can be highly dependent on the population used in training sets–a representative sample of the general population should be used in the development of this technology to ensure results are broadly applicable. Further, computer-aided detection (CAD), an earlier computer system aimed at improving mammography interpretation, showed great promise in experimental testing but fell short in real-world settings.

While there is promise for AI in healthcare, we would do well to consider and extensively test AI systems before we consider deploying them. The sheer volume of data needed to develop and test AI algorithms for clinical tasks is itself a huge hurdle, while ensuring that our algorithms aren’t biased by the sample population is also vitally important. To develop and wisely use such AI systems, we must pay attention to patient privacy, how data are stored and used, by whom, and with what type of oversight.

Advances & Business

Concerns & Hype

Analysis & Policy

Expert Opinions & Discussion within the field

  • Top minds in machine learning predict where AI is going in 2020 - AI is no longer poised to change the world someday; it’s changing the world now. As we begin a new year and decade, VentureBeat turned to some of the keenest minds in AI to revisit progress made in 2019 and look ahead to how machine learning will mature in 2020.

  • Ethics of Technology Needs More Political Philosophy - The two approaches of answering questions about the ethics of self-driving cars are both lacking. We should neither deduce answers from individual moral theories nor should we expect social science to give us complete answers. In this article, Himmelreich argues that we should turn to political philosophy in addition to moral philosophy.

That’s all for this week! If you are not subscribed and liked this, feel free to subscribe below!

More like this
Follow us
Get more AI coverage in your email inbox: Subscribe