Connect with us

Artificial Intelligence

Labeling and Rating Potentially Harmful AI Systems Is Inherently Complex, Say Brookings Panelists

Jericho Casper

Published

on

Photo of Mark MacCarthy by the Federal Trade Commission from November 2018

October 5, 2020 — Germany has attempted to improve awareness of the potential degree of consumer harm that artificial intelligence systems cause by developing categories about the degree of environmental, economic, and societal impact it could cause.

To hear expert opinions on whether policymakers and stakeholders in the United States should adopt similar models and utilize AI certifications, ratings and labeling, the Brookings Institute hosted a panel on Thursday to consider potential “blind spots.”

“We need some sort of testing or auditing mechanism,” said Elham Tabassi, chief of staff of the information technology laboratory at the National Institute of Standards and Technology.

“Standards for AI are lacking right now,” Tabassi said, and experts “don’t know how to test AI systems for bias.”

Mark MacCarthy, adjunct faculty of communication, culture, and technology at Georgetown University, detailed what has traditionally occurred when industry applied ethics systems in an attempt to manage new forms of media or technology.

The Platform for Privacy Preferences Project, or P3P, was “a system developed when the public was worried about safety online, which allowed users to opt out of online tracking,” said MacMarthy.

Screenshot from the Brookings Institution webinar

The idea was that websites would post their privacy policies in P3P format and web browsers would download them automatically and compare them with each user’s privacy settings.

In the event that a privacy policy did not match the user’s settings, the browser could alert the user, block cookies, or take other actions automatically. Yet, the system “did not come to anything.”

“A vast number of systems have been proposed for ethics of artificial intelligence,” said MacMarthy. “The rating system is not well developed, but I hope this conversation can help it,” he said.

MacCarthy maintained that he held two reservations. First, he noted that most non-mandatory ethics regulations have failed because of a lack of industry buy-in. It “would need to be mandatory,” as “industry has no reason to use it,” said MacCarthy.

Second, “even with coercion the objective of an AI rating system is to give consumers more information about these systems, and to be honest, that may not be enough,” he concluded.

“There are so many complications in the AI process, it makes it extremely difficult,” to pinpoint where systems go wrong, said John Villasenor, nonresident senior fellow of governance studies at the Center for Technology.

One issue, according to Villasenor, is that AI systems are “produced by commercial entities, to make a profit.”

“Many companies don’t put their source code on the internet because they want to remain competitive,” he said. Further, there are over “500,000 lines of code, it may not be easy to figure out exactly what the algorithm is doing.”

The panelists did not rally around the German approach of looking at high risk algorithms and labeling them.

“How to decide if something is high risk?,” questioned Elham. He also wondered if labeling and categorizing AI would create a false sense of complacency.

Assistant Editor Jericho Casper graduated from the University of Virginia studying media policy. She grew up in Newport News in an area heavily impacted by the digital divide. She has a passion for universal access and a vendetta against anyone who stands in the way of her getting better broadband.

Artificial Intelligence

Staying Ahead On Artificial Intelligence Requires International Cooperation

Benjamin Kahn

Published

on

Screenshot from the webinar

March 4, 2021—Artificial intelligence is present in most facets of American digital life, but experts are in a constant race to identify and address potential dangers before they impact consumers.

From making a simple search on Google to listening to music on Spotify to streaming Tiger King on Netflix, AI is everywhere. Predictive algorithms learn from a consumer’s viewing habits and attempt to direct consumers to other content an algorithm thinks a consumer will be interested in.

While this can be extremely convenient for consumers, it also raises many concerns.

Jaisha Wray, associate administrator for international affairs at the National Telecommunications and Information Administration, was a panelist at a conference hosted Tuesday by the Federal Communications Bar Association.

Wray identified three key areas of interest that are at the forefront of AI policy: content moderation, algorithm transparency, and the establishment of common-ground policies between foreign governments.

In addition to all the aforementioned uses for AI, it also has proven to be an indispensable tool for websites like Facebook, Alphabet’s Youtube, and myriad other social media platforms in auto-moderating their content. While most social media platforms employ humans to review various decisions made by AI (such as Facebook’s Oversight Board), most content is first handled by AI moderators.

According to Tubefilter, in 2019 more than 500 hours of video content were uploaded to Youtube every minute; in less than 20 minutes, a year’s worth of content is uploaded.

Content moderation, algorithm transparency, foreign alignment

On this scale, AI is necessary to police the website, even if it not a perfect system. “[AI] is like a thread that’s woven into every issue that we work on and every venue,” Wray explained. She described how both governments and private entities have looked to AI to not only moderate somewhat mundane things such copyright issues, but also national security issues like violent extremist content.

Her second point pertained to algorithm transparency. She outlined how entities outside of the U.S. have sought to address this concern by providing consumers with the opportunity to have their content reviewed by humans before a final decision is made. Wray pointed to the European General Protection Regulation, “which enshrines the principle that every person has the right not to be subject to a decision solely based on automated processing.”

Her final point raised the issue of coordinating these efforts between different international jurisdictions—namely the U.S. and its allies. “We’re really trying to hone-in on where our values align and where we can find common ground.” She added that coordination does not end with allies, however, and that it is key that the U.S. also coordinate with authoritarian regimes, allied or otherwise.

She said that the primary task facing the U.S. right now is simply trying to determine which issues are worth prioritizing when it comes to coordinating with foreign governments—whether that is addressing the spread of AI, how to police AI multilaterally, or how to address the use of AI by adversarial authoritarian regimes.

Technology needs to be built with security in mind

One of Wray’s co-panelists, Evelyn Remaley, who is the associate administrator for the NTIA’s Office of Policy Analysis and Development, said all multilateral cybersecurity efforts related to AI must be approached from a position of what she called a “zero-trust model.” She explained that this model operates from the presupposition that technology should not and cannot be trusted.

“We have to build in controls and standards from the bottom-up to make sure that we are building in the security layer by layer,” Remaley said. “It’s really that premise of ensuring that we realize that we’re always going to have vulnerabilities within this technical development space.”

Remaley said that increasing competition and collaboration can only be safely achieved with a zero-trust mindset.

Continue Reading

Artificial Intelligence

Connectivity Will Need To Keep Up With The Advent Of New Tech, Says Expert

Samuel Triginelli

Published

on

Screenshot from the webinar

February 24, 2021 – It used to be that technology had to keep up with the deployment of the growing ubiquity of broadband innovations. But the pace of technological advancements in the home is starting a conversation about whether connectivity can keep up.

That’s according to Shawn DuBravac, an accountant and author of a book about how big data will transform our everyday lives, who argues that the pandemic has illustrated the need for broader connections in the home to meet the need of future technologies. He was speaking on Tuesday at the conference of NTCA – The Rural Broadband Association.

Emerging consumer technologies, such as Samsung’s robots, which will perform tasks including loading a dishwasher, serving wine, and setting a dinner table, are redefining the conversation about how connectivity at home will manage them, DuBravac argues.

Health companies are also introducing “companion robots” focused on interacting with seniors. With its artificial intelligence and sensors, these robots develop a personality to adapt to the needs of consumers so social distancing does not become a disadvantage for care.

As such, the pandemic has grown the telehealth industry. With more people avoiding going to hospitals, the creation of watches, belts, scales that are connected to share information with medical professionals is further requiring better broadband connectivity to keep up.

But it’s not like the industry isn’t paying attention. Mesh network technologies, which utilize multiple router-like devices to enhance coverage inside the home, have started to emerge just as smart-home technologies illustrated the need for broader connectivity that better enhanced coverage as Wi-Fi signals experienced degradation through walls.

Continue Reading

Artificial Intelligence

AI the Most Important Change in Health Care Since Introduction of the MRI, Say Experts

Samuel Triginelli

Published

on

Screenshot from the webinar

February 7, 2021 – Artificial Intelligence is the most important technological change in health care since the introduction of the MRI, experts said at a Thursday panel discussion about European tech sponsored by the Information Technology and Innovation Foundation.

AI will not be replacing doctors and nurses, but empowering decision-maker with new resources, according to those participating in the discussion on “How Can Europe Enhance the Benefits of AI-Enabled Health Care?”

For example, pharmaceutical companies are using AI for the speedy development of vaccines, panelists said. Additionally, AI is helping address the uneven ratio of skilled doctors to patients, assist health-care professionals in complex procedures, and deliver personalized health care to patients.

Yet, for AI technologies to reach their potential, European Union actors need to create regulations governing transparency, they said.

How AI works in healthcare

AI works through big collections of data that validate algorithms. These help explain certain solutions and detect anomalies in the data set of patients.

But algorithm-creation needs to be held to higher standards than they are currently. Systemic errors can easily enter in on a large scale, said Elmar Kotter, chairperson of the eHealth and Informatics Subcommittee of the European Society of Radiologists.

AI should have been used more during the early stages of the COVID-19 pandemic, said Maria Manuel Marques, on the Special Committee on Artificial Intelligence in a Digital Age.

AI helps treat more patients at a faster rate, and with consistency and agility, said Chris Walker, chair of the working group on digital health for the European Federation of Pharmaceutical Industries and Associations. It helps provide new insights and improve treatment by allowing early-stage treatment of diseases.

Europe faces great challenges because of people’s misconception of what AI can do, panelists said. It is not to replacing doctors and nurses, but empowering with decision-making resources.

More trust would come if companies would conduct safe experimentation by testing and showing examples of how AI can improve the life of health care workers and patients, said Marques.

Regulations of data is crucial for hospitals to trust the products. Moreover, patients must have privacy with their information. Regulations will help them understand what’s been done in the manufacture of AI system, and to what use data will be put.

Ander Elustondo Jauregui, policy officer for Digital Health, added that data quality was an important indicator of the maturity of an AI system. That providing assurances for doctors, he said.

Continue Reading

Recent

Signup for Broadband Breakfast

Get twice-weekly Breakfast Media news alerts.
* = required field

Trending