This week in AI and Machine Learning: Looking ahead to Google Cloud Next, the AI Bill of Rights, and more.
Author’s Note
From October 11th-13th, next week’s Google Cloud Next conference will connect computing and technology professionals from across the globe for three days of high-impact keynotes and panel discussions.
Among this year’s presenters is Plainsight’s Co-Founder and CEO, Carlos Anchia. Carlos joins two Google Cloud Product Managers, Fabien Blanc-Paques and Nelson Gonzales, and Brain Corp Vice President, Botond Szarmáry to discuss and demonstrate the transformative power of repeatable enterprise computer vision in a session titled, Deliver next generation business insights with computer vision applications.
Attendees can expect to learn about how Plainsight and Google Cloud are working together to help businesses:
- Accelerate speed to value of computer vision with transformative insights
- Productionalize successful end-to-end computer vision across diverse use cases
- Enable cost reduction, resource efficiency, as well as responsible computer vision management and oversight.
That’s not the only session we’re looking forward to next week. Next registrants can prepare by organizing a playlist of the presentations that look most appealing to them. Check out Carlos’ complete list of must-attend sessions and register for Google Cloud Next today to start building a playlist of your own.
Artificial Intelligence News
White House Announces AI Bill of Rights
On Tuesday, the Biden administration published a whitepaper entitled Blueprint for an AI Bill of Rights, intended to draw attention to issues related to AI and kickstart a federal response. According to the five key principles identified in the document, Americans in the AI-age have the following rights: protection against unsafe and ineffective systems, protection from discrimination by systems and algorithms, protections from abuse data practices and agency over how data is used; information related to which systems are used as well as how and why they work; and the option to opt out of automated support where appropriate.
Though the tech world is still digesting the news, some early responses have been less than pleased. Khari Johnson, a senior writer for WIRED, argues that the new AI Bill of Rights does not go nearly far enough to reign in Big Tech and protect consumers. “Unlike the US Bill of Rights,” he notes, “the AI version will not have the force of law.” He compares the blueprint unfavorably to similar corporate and government charters. While the principles and goals discussed “are usually directionally right,” he writes, such initiatives are ultimately too vague.
Khari quotes from numerous experts who agree the blueprint should go further. Annette Zimmerman, an AI ethics researcher, acknowledges that she’s both impressed and underwhelmed with the principles outlined by the White House Office of Science and Technology (OSTP). Though the five central tenets cover a lot of ground, they do not take into account that AI cannot always be corrected with AI. “We can’t articulate a bill of rights,” she remarks, “without considering non-deployment.” OSTP officials have characterized the document as more a call to action than a framework for enforcing its recommendations.
Avoiding Unnecessary Biopsies with AI
When a biopsy comes back negative, it’s a mixed blessing of sorts. While every patient is glad to learn they’re cancer free, there is potential frustration in learning you’ve agreed to invasive procedures unnecessarily. In many cases, according to New York University postdoctoral fellow Jan Witowski, radiologists “err on the side of, let’s just biopsy a lot of people” because they lack the data necessary to make a confident decision either way. This ‘better safe than sorry’ approach could mean unnecessary pain, stress, discomfort, and expense for patients. According to STAT, two to four biopsies return benign results for every one sample that shows cancer cells.
Witowski and colleagues from both NYU and Poland’s Jagiellonian University have developed an AI tool to predict the likelihood of breast cancer diagnosis from MRI scans and expert recommendations. The model, discussed in a recent paper for Science Translational Medicine, was trained on more than 20,000 labeled breast scans and proved successful across a broad range of patient demographics. Its diagnostic prowess has reduced the likelihood of false positives and could mean fewer patients forced to undergo unnecessary procedures. Next up for Witowski and his team are clinical trials.
Meet Florence, WHO’s AI Healthcare Pro
In California last week, Tesla’s annual AI Day included the debut of a new humanoid robot prototype. Far away in Qatar, this week’s World Innovation Summit for Health (WISH) has seen the World Health Organization (WHO) reveal the beta version of Florence version 2.0. Initially deployed to help fight the spread of COVID-19 misinformation, this AI-powered healthcare assistant. will interact with scientists and public health officials throughout WISH to showcase new capabilities.
The new and improved digital healthcare worker can now converse in five languages, with two more to follow. In addition to providing information on COVID-19 symptoms, vaccines, and prevention, the Soul Machines’ creation has helped dispense guidance on mental health care and smoking cessation. Florence is a key result of WHO’s ongoing Sport for Health partnership with the Qatar Ministry of Public Health. The partners intend to use the nation’s upcoming FIFA World Cup as a stage for showcasing Florence and discussing other high-priority health and safety initiatives
Join our Community
See you next week! Until then, keep the conversation going on Plainsight’s AI Slack Channel.
About the Author & Plainsight
Bennett Glace is a B2B technology content writer and cinephile from Philadelphia. He helps Plainsight in its mission to make vision AI accessible to entire enterprise teams.
Plainsight’s vision AI platform streamlines and optimizes the full computer vision lifecycle. From project strategy, through model deployment, and ongoing monitoring, Plainsight helps customers successfully create and operationalize vision AI applications to solve highly diverse business challenges.
