Hello!
Welcome to Hold The Code. This week, for our tenth edition, we have many reasons to celebrate.
RAISO hosted its first-ever speaker event last week with Juyoun Han, an AI fairness and data privacy lawyer. Thank you very much to all of our readers who attended and made it such an insightful event. Stay tuned for more speakers and educational workshops this quarter!
Additionally, please welcome Hold The Code's newest writer, Sophie Lamb! Sophie is a junior at Northwestern University studying journalism and political science. We're looking forward to sharing more perspectives on all things AI and contemporary tech.
So without further ado...
Can Netflix Algorithms Cure Alzheimer's Disease?
Researchers at the University of Cambridge recently published a study suggesting that the same machine-learning technology used in apps like Netflix and Facebook can have powerful applications to neurodegenerative research. They argue that AI can predict the biological language of Alzheimer's and other diseases, enabling scientists to understand and correct grammatical mistakes inside cells.
First: A little biology lesson
(Put your STEM cap on.)
The human body is home to thousands and thousands of proteins, and scientists don't yet know the function of many of them.
In Alzheimer's disease, which affects 50 million people worldwide, proteins go rogue, forming clumps and killing healthy nerve cell
Understanding these things called bimolecular condensates — droplets of proteins found in cells — is necessary for Alzheimer's research. Scientists believe that defects connected with the droplets can lead to diseases.
How can AI help?
You know on Netflix — how it miraculously recommends shows that are precisely what you're looking for? That's the exact type of algorithm Dr. Saar and his team at Cambridge pursued when conducting their Alzheimer's research.
They built a language-processing program that took in data on known proteins. "In the same way apps like WhatsApp know how to suggest words for you to use," their algorithm learned and predicted the language of these proteins.
This enabled the researchers to ask the program-specific questions about the grammar that leads some cells to form proteins that condensate inside cells, causing disease.
Understanding this perplexing feature of neurodegenerative diseases can help unlock profound insights and potentially pave the path for a future solution.
When Your Substitute Teacher is a Robot...
The use of AI in education has become a highly discussed - and somewhat controversial - topic. According to Stanford University’s 2021 AI Index, more than $40 billion was invested into AI startups in 2020, and both experts in AI and in education expect these technologies to be increasingly integrated into classrooms.
What can these systems do?
In short: a lot. These systems can range from AI-powered teaching assistants that automatically grade assignments and provide real-time feedback to students to robots that help teachers in organizing socialization in their classrooms.
Joanna Smith, the founder of ed-tech company AllHere, recently launched new product features that include an AI chatbot that checks in with students who are missing class regularly, an issue that has been exacerbated in many schools by the COVID-19 pandemic. This system can even provide families with health care referrals or with support in solving computer-related issues.
What's the controversy?
Some experts advise caution when integrating AI with education. One issue is that research goals are relatively future-oriented and may not address the practical, daily issues that education professionals regularly experience. In addition, concerns surrounding biased datasets and increased surveillance exist in this space.
When developing any AI system that will impact the education sphere, it is crucial that a human-centered approach is used to design, build, and improve these systems.
Jeremy Roschelle, a principal investigator at the Center for Integrative Research in Computing and Learning, puts it best by stating we need to “start from what is good for teaching and learning...and not from what AI can do for me.”
Amazon Drivers Face AI Micromanagement
Amazon is requiring their drivers to sign a “biometric consent” form to allow AI cameras into their vehicles to capture photographs, vehicle data, and “potentially risky driver behavior, such as distracted driving or drowsy driving.”
Biometrics: physical or human characteristics used for authentication purposes (think fingerprints and eye scans)
What data will be collected?
The technology provided by Netradyne will record facial recognition and other biometric data to confirm the identity of drivers and track their behavior during deliveries, such as yawning and texting while driving. The AI system will also record vehicles’ location, acceleration, braking, and following distance to recognize potential traffic violations. If drivers refuse to comply, they risk unemployment.
What's the impact?
As Amazon claims the cameras will increase community safety, some drivers believe the company is going a step too far.
Speaking to Reuters, 22-year-old driver Henry Search calls it an “invasion of privacy.”
“We are out here working all day, trying our best already,” Search said. “The cameras are just another way to control us.”
Amazon currently employs 75,000 drivers among 800 companies. Where do you stand?
Weekly Feature: "How Blockchain Can Simplify Partnerships"
Our current ~~ work from home era ~~ has emphasized the need for better forms of virtual communication and collaboration. A recently published piece in the Harvard Business Review argues that blockchain technology can offer exciting solutions.
Some terminology
Blockchain seems complicated, and it definitely can be, but its core concept is really quite simple. A blockchain is a type of database.
So, what's a database, really?
Answer: a collection of information that is stored electronically on a computer system.
Databases can be spreadsheets used by one person or a small group of people. But they can also be much larger — and housed on thousands of powerful computers.
Yet, blockchains differ from typical databases.
Blockchains collect information in groups, AKA: blocks.
Blocks have certain storage capacities. Once they're full, a new block is formed and they're linked together in a — wait for it — blockchain.
Significantly, this process creates an irreversible timeline of data; each block in the chain is given an exact timestamp.
And one more important thing to know:
Blockchains are considered decentralized databases.
A traditional database might be stored on a server across 10,000 computers all housed under one roof in a company warehouse.
However, blockchains can be implemented across thousands of different computers across varying geographic locations. The computers that make up a blockchain network are often called "nodes."
OK, so how can businesses use them?
The Harvard Business Review argues that blockchains can simplify partnerships by creating a reliable record of transactions, avoiding costly disputes, and changing how deals are made.
1). Blockchains are "digital ledgers" — meaning several people have joint control over the shared information. This makes them ideal for situations where trust and information sharing are important.
2). Blockchains can be paired with "smart contracts": programmed codes that are automatically executed once certain conditions are met.
For instance, companies like Maersk, the world’s leading logistic company, have already begun using smart contracts to automate a variety of tasks in the shipping process. Their program increases efficiency because execution is automated, and the enforcement of the shipping agreement relies on neither the court nor social sanctions, but rather on a set of protocols representing a self-contained and autonomous system of rules.
Moreover, with blockchain, "Information becomes secure, immutable, transparent, and traceable when stored and processed on the blockchain. With a much higher level of accuracy and authentication, the parties have less concern about possible collaboration failures."
Read the full Harvard Business Review piece here, and for a more technical description of blockchains, read this article.
Written by Lex Verb, Molly Pribble, and Sophie Lamb