AI News Roundup — April 2020
by Gabriella Runnels and Macon McLean
The AI News Roundup provides you with our take on the coolest and most interesting Artificial Intelligence (AI) news and developments each month. Stay tuned and feel free to comment with any stories you think we missed!
_________________________________________________________________
AI Is One Hell of a Drug
Defeating the COVID-19 pandemic will require the combined forces of the world’s greatest scientific minds, both human and artificial alike. British startup BenevolentAI has put its artificial intelligence research to work, and its AI — a “search engine on steroids” that uses pharmaceutical data and research findings to identify potential drug applications — has made a promising discovery.
Recent results indicate that experts might be able to repurpose a rheumatoid arthritis medication to fight the coronavirus. By attacking the “cellular machinery” that the virus uses to infect the host, it’s possible that this medication could serve as an effective treatment for COVID-19. Drug trials are set to start in the US this month, with initial results coming as early as June of this year.
The White (Noise) Whale
Whales are big, but the ocean is much, much bigger. As it turns out, finding a whale in the ocean is a pretty daunting task, especially as whale populations decline and relocate as a result of climate change, making it hard for scientists to track and study them.
In order to better track humpback whales, NOAA researcher Ann Allen partnered with Google to develop a machine learning algorithm that differentiates the whales’ unique songs from the cacophony of other ocean sounds. The neural network they created, similar to one used by YouTube for identifying sounds in videos, can help scientists identify and protect these vulnerable creatures.
Status: It’s Complicated
If you’ve been following coronavirus news, you may have noticed some pretty wide-ranging estimates of the total predicted COVID-19 fatalities. Some sources show only a few hundred thousand casualties, while others are forecasting figures in the millions. What’s going on here?
Predicting coronavirus fatalities might seem like a straightforward function of the number of people who’ll contract the disease and the rate at which those people will survive it. Unfortunately, there’s almost nothing simple about actually arriving at accurate numbers. For a breakdown of the complexities of modeling COVID-19, check out this great post from the folks over at FiveThirtyEight.
Bias Bounties for Ethics in AI
Another day, another set of AI ethics guidelines/tools from major corporations and research institutions. But this one has a fun idea: encourage developers, machine learning engineers, data scientists, and the like to eradicate bias from AI by paying them to do so.
In the same way that “bug bounties” reward software engineers for finding and/or fixing bugs in code, these new “bias bounties” follow from the simple economic principle of rewarding behaviors that should be encouraged. The paper goes on to suggest that these incentives could extend beyond removing or reducing bias to further preserving privacy, increasing interpretability, or improving other aspects of an algorithm/AI system.
AI Contact Tracing App Garners Mixed Reactions
A new app developed by the Indian government known as Aarogya Setu became the first app ever to reach 50 million users in 13 days or fewer. This flurry of activity came at the behest of India’s Prime Minister Narendra Modi, who strongly encouraged citizens to use it in the fight against COVID-19.
The app performs a kind of digital contact tracing, which involves evaluating a phone user’s history (often including location data) to evaluate whether or not he or she was likely to have been in contact with a known carrier of the virus. While some support the app’s mission, others are concerned that user too readily relinquish privacy for untested security measures.
That’s it for this month! In case you missed it, here’s last month’s roundup with even more cool AI news. Check back in May for more of the most interesting developments in the AI community (from our point of view, of course).
_________________________________________________________________
If you liked this blog post, check out more of our work, follow us on social media (Twitter, LinkedIn, and Facebook), or join us for our free monthly Academy webinars.
