News: Center for Advancing Safety of Machine Intelligence

Parents Put Trust in Waymo’s Self-Driving Cars for Safer School Runs

September 2024 Notes on the News: CASMI Director Kristian Hammond explains why parents are trusting Waymo's self-driving cars to take their kids to school.

The Complex Promise and Perils of AI in Policing

September 2024 Notes on the News: CASMI Director Kristian Hammond discusses the use of AI in Policing.

Degenerative AI: The Risks of Training Systems on their own Data

September 2024 Notes on the News: CASMI Director Kristian Hammond explain the risks of training systems on their own data.

The AI Summarization Dilemma: When Good Enough Isn’t Enough

August 2024 Notes on the News: CASMI Director Kristian Hammond explains the issue of tasking AI with summarizing text.

The Hallucination Problem: A Feature, Not a Bug

August 2024 Notes on the News: CASMI Director Kristian Hammond discusses the role of hallucinations in large language models.

A Step Forward with the Kids Online Safety Act—but with Caution

August 2024 Notes on the News: CASMI Director Kristian Hammond responds to legislation that Congress is working to pass, focused on online safety for children.

Time to Roll Up Your Sleeves: The Real Work for Generative AI Begins

August 2024 Notes on the News: A recent report says that the hype surrounding generative AI is waning. This means it's now time to put in real work, according to CASMI Director Kristian Hammond.

Trump's Deepfake Strategy: Eroding Truth with a World of Lies

August 2024 Notes on the News: CASMI Director Kristian Hammond discusses former President Donald Trump's use of deepfakes.

OpenAI’s Strawberry: A Step Towards Advanced Reasoning in AI

August 2024 Notes on the News: CASMI Director Kristian Hammond reacts to reports that OpenAI is working on a new reasoning technology called Strawberry.

Misinformation at Scale: Elon Musk's Grok and the Battle for Truth

August 2024 Notes on the News: CASMI Director Kristian Hammond addresses an open letter that five Secretaries of State penned to Elon Musk, urging the owner of X to immediately implement changes to his social media platform's AI search assistant, Grok, to "ensure voters have accurate information in this critical election year.”

OpenAI: The Battle over Search

August 2024 Notes on the News: CASMI Director Kristian Hammond reacts to OpenAI's announcement that it is testing a new search engine.

The Peril of Perfection: Google's Gemini Ad and the Cost of Thoughtfulness

August 2024 Notes on the News: CASMI Director Kristian Hammond responds to a Google Gemini commercial which was advertised during the Olympics before being pulled due to backlash.

AI, Math, and Language: Bridging the Divide

August 2024 Notes on the News: CASMI Director Kristian Hammond explains why large language models are bad at math.

Testing Large Language Models

July 2024 Notes on the News: CASMI Director Kristian Hammond argues that our current approach of using benchmarks to judge AI tools seems to be flawed.

AI and Education

July 2024 Notes on the News: CASMI Director Kristian Hammond reflects on a conversation had at a dinner on the use of AI in education.

A.I. Can Write Poetry, but It Struggles With Math

July 2024 “The A.I. chatbots have difficulty with math because they were never designed to do it,” CASMI Director Kristian Hammond told The New York Times.

from The New York Times

Artificial Intelligence's Role in Law

July 2024 Notes on the News: CASMI Director Kristian Hammond comments on the 18th Edition of the International Conference on Substantive Technology in Legal Education and Practice, or SubTech, an annual event which examines the impact of technology in law.

Microsoft and Apple Won't Join OpenAI Board Seats: A Setback for AI Progress

July 2024 Notes on the News: CASMI Director Kristian Hammond reacts to Microsoft and Apple's decision not to join OpenAI's observer board.

OpenAI’s Quest for Artificial General Intelligence: Vision or Mirage?

July 2024 Notes on the News: CASMI Director Kristian Hammond reacts to OpenAI's five-step plan toward artificial general intelligence.

The AI Dilemma: Google's Greenhouse Gas Emissions Surge

July 2024 Notes on the News: CASMI Director Kristian Hammond responds to Google's latest environmental report, which shows a 48% increase in greenhouse gas emissions in the last five years.

Apple's Pragmatic AI: A Smarter Approach to Everyday Integration

June 2024 Notes on the News: CASMI Director Kristian Hammond comments on Apple's strategy to incorporate AI into its products.

The Open Letter: A Disappointing Focus on Legalities Over Critical Issues

June 2024 Notes on the News: CASMI Director Kristian Hammond reacts to an open letter from former employees of OpenAI, who are warning about the risks posed by the technologies and the need to protect whistleblowers.

First Came ‘Spam.’ Now, With A.I., We’ve Got ‘Slop’

June 2024 The term slop became more prevalent last month when Google incorporated its Gemini A.I. model into its U.S.-based search results. That created a problem, CASMI Director Kristian Hammond told The New York Times.

from The New York Times

Reflecting on Blockbuster as We Consider the Future of Work

June 2024 Notes on the News: CASMI Director Kristian Hammond shares a touching story about Blockbuster to highlight the potential of AI's role in the future of work.

Trust is Not a Marketing Issue

June 2024 Notes on the News: CASMI Director Kristian Hammond comments on the dangers of marketing friendly and cute AI systems.

The Risk of Google’s Shift from Search Engine to Answer Machine

June 2024 Notes on the News: CASMI Director Kristian Hammond comments on Google's shift from search engine to answer machine with its decision to integrate AI.

The New Dawn of AI Evaluation: NIST's ARIA

May 2024 Notes on the News: CASMI Director Kristian Hammond comments on the latest program from the National Institute of Standards and Technology (NIST): ARIA — Assessing Risks and Impacts of AI.

Navigating the Library of Babel in the Age of AI

May 2024 NOTES ON THE NEWS: As we move deeper into the age of automatically generated content, we find ourselves edging closer to a modern-day Library of Babel.

OpenAI's Misstep with Scarlett Johansson's Voice: A Case Study in Tone Deafness

May 2024 Notes on the News: CASMI Director Kristian Hammond addresses OpenAI's decision to use a voice similar to Scarlett Johansson's for their latest model, GPT-4o.

Funding AI: Balancing Speed and Safety

May 2024 Notes on the News: CASMI Director Kristian Hammond reacts to a bipartisan bill that would offer massive funding for artificial intelligence.

What Google AI means for you — and your search results

May 2024 Google’s idea is a great one, but needs further validation, according to CASMI Director Kristian Hammond.

from Northwestern Now

The Challenges of Regulating AI and the Role of Behavioral Science

May 2024 CASMI Director Kristian Hammond participated in a panel at the Behavioral Science & Policy Association's annual conference to discuss the role behavioral science can play in regulating AI.

from Behavioral Scientist

Computer Scientist Shares Enlightening Journey to See the ‘Value in Expression’

May 2024 Haoqi Zhang wants other computer scientists, including researchers in artificial intelligence (AI) and human-computer interaction (HCI), to examine their thinking from a consequentialist point of view (which judges actions solely by their consequences, i.e. goal-reaching methods) to also incorporate a dialectical mindset (which emphasizes the inherent value of an activity, e.g. being a good friend).

Meta's AI Chatbot Causing Some Confusion

May 2024 Notes on the News: CASMI Director Kristian Hammond addresses recent hallucinations from Meta's AI chatbot, which confused some Facebook users.

Suite of Approaches That Can Be Applied to Uncover Possible Harm

May 2024 Notes on the News: CASMI Director Kristian Hammond proposes a suite of approaches that can be applied to uncover possible harm.

Anthropic CEO Comments on the AI Arms Race

April 2024 Notes on the News: CASMI Director Kristian Hammond comments on the AI arms race.

How Educators and Students Should Use Language Models

April 2024 Notes on the News: CASMI Director Kristian Hammond explains how educators and students should use language models.

Use of AI in War

April 2024 Notes on the News: CASMI Director Kristian Hammond reacts to recent reports of AI use during warfare.

Research Explores How a Wikipedia-Like Approach Could Improve AI Evaluation

April 2024 CASMI researchers developed a tool that supports community-driven data curation for AI evaluation.

A New Era of AI-Generated Content

April 2024 MSAI director Kristian Hammond discusses OpenAI's new Sora model and its ability to create realistic video.

from Master of Science in Artificial Intellience (MSAI)

Pornographic Deepfakes in Schools

April 2024 Notes on the News: CASMI Director Kristian Hammond addresses the pressing issue of public school students creating pornographic deepfakes of their classmates.

Pennsylvania AI Buyer Beware Bill

April 2024 Notes on the News: CASMI Director Kristian Hammond reacts to a proposal in Pennsylvania that would require you to tag content that is generated by a machine.

Meta's Nick Clegg Downplaying AI Election Fears

April 2024 Notes on the News: CASMI Director Kristian Hammond responds to an attempt from Meta’s President of Global Affairs Nick Clegg at calming people's fears regarding the use of generative AI to influence people’s political views.

Adam Schiff AI Copyright Bill

April 2024 Notes on the News: CASMI Director Kristian Hammond reacts to a proposal which would require companies to disclose their use of copyrighted work for training generative AI models.

How Companies May Be 'Acting Irresponsibly' When Training AI Models

April 2024 Notes on the News: CASMI Director Kristian Hammond reacts to a report which claims tech companies that are building large language models or other foundation models have perhaps been playing a little fast and loose with copyright law and user agreements.

China's Use of AI to Create Propaganda in Elections

April 2024 Notes on the News: CASMI Director Kristian Hammond addresses a report that China is using AI to create propaganda in the US, South Korea, and India.

EU-US Collaboration on AI Safety

April 2024 Notes on the News: CASMI Director Kristian Hammond reacts to an announcement that the EU and US will work together to evaluate artificial intelligence models.

Musical Artists' Warning about AI Use

April 2024 Notes on the News: CASMI Director Kristian Hammond addresses an open letter from musical artists which urges tech platforms to stop devaluing music.

Online Users Navigate Mental Health Information, Empowered to Help Make Changes

March 2024 Access to mental health services is limited, so people are turning to the internet to find the information they need. New research finds this information-gathering process is cyclical and involves both search engines and social media platforms. The discovery is aimed at empowering online users to help make design changes.

Engaging with Lawmakers, Business Leaders to Promote AI Safety

March 2024 CASMI is engaging with lawmakers and business leaders as it educates the public about the impact of artificial intelligence (AI) technologies.

Opinion: The City that Works needs to put AI to work

March 2024 CASMI Director Kristian Hammond explains how the city of Chicago should embrace its Second City nickname to put AI to work.

from Crain's Chicago Business

Harnessing the Power of Data Science and Artificial Intelligence

March 2024 CASMI Director Kristian Hammond is a member of Northwestern's newly created Data Science and Artificial Intelligence (DS/AI) Steering Committee.

from Northwestern Leadership Notes

CASMI Recognizes Research Focused on Preventing, Mitigating AI Harms

February 2024 CASMI is supporting AI safety research through the AI Incidents and Best Practices track at the Thirty-Sixth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-24).

Ultra-realistic artificial intelligence video is here

February 2024 CASMI Director Kristian Hammond tells CNN, "The clarity of truth we thought we had with recorded photography and video is gone."

from CNN

OpenAI video-generator Sora risks fueling propaganda and bias, experts say

February 2024 CASMI Director Kristian Hammond says AI products carry the potential to improve and ease video storytelling, but they could also supercharge internet misinformation and enhance government propaganda, blurring the already-faint line between real and fake content online.

from ABC News

AI professional headshots are quick and easy. But should you use one?

February 2024 “It’s time to admit we’re soon going to be surrounded by representations of the world which aren’t real … and they will become less and less real every day,” said CASMI Director Kristian Hammond.

from Washington Post

Oversight Board Urges Meta to Rethink its Policy on Manipulated Media in High-Stakes Election Year

February 2024 CASMI Director Kristian Hammond joined “Chicago Tonight” on Feb. 5, 2024, to discuss the implications of AI on the election.

from WTTW

Tracking Political Deepfakes: New Database Aims to Inform, Inspire Policy Solutions

January 2024 The creators of the Political Deepfakes Incidents Database ‒ Christina Walker, Purdue University PhD candidate in political science; Daniel Schiff, Purdue assistant professor of technology policy; and Kaylyn Jackson Schiff, Purdue assistant professor of political science ‒ won the inaugural Northwestern Center for Advancing Safety of Machine Intelligence (CASMI) AI Incidents and Best Practices Paper Award.

The AI Will See You Now

January 2024 If AI is not already in your doctor’s office, you can bet it’s coming soon — with the potential to create major improvements in our health care and well-being.

from Northwestern Magazine

AI and the Law: The Jury's Still Out

December 2023 In law and legal services, large language models (LLMs) — such as OpenAI's ChatGPT and Google's Bard — could be a game-changer. Yet these technologies pose more questions than answers.

from Northwestern Engineering

Shirley Ryan’s Learning for Life series dives into artificial intelligence

December 2023 The popular lecture series drew record numbers with this year’s topic. CASMI Director Kristian Hammond was the series’ first featured speaker.

from Northwestern Now

Empowering Human Knowledge for More Effective AI-Assisted Decision-Making

November 2023 CASMI researchers from Carnegie Mellon University traveled on Nov. 6-9 to Delft, Netherlands to present findings from their research project, which is exploring key differences between human and AI judgment to support more effective human-AI decision-making.

Focusing on advancing safety and sustainability, UL grows its partnership with Northwestern

November 2023 In a cooperative initiative, UL Research Institutes (ULRI) and Northwestern University are working to bring safety to the forefront of the rapidly expanding field of AI and machine learning through their leadership of the Center for Advancing Safety of Machine Intelligence (CASMI) research hub.

from Corporate Engagement

Addressing Gender and Intersectional Bias in Artificial Intelligence

November 2023 On November 10, the Gender Equity Initiative at Northwestern Pritzker School of Law and the Northwestern University Law and Technology Initiative cohosted the “Gender and Intersectional Bias in Artificial Intelligence Conference” to discuss how and why AI reflects and exacerbates historical and social inequities and innovative technical and legal approaches to mitigate those biases.

from Computer Science Department

CASMI Attends Chicago Innovation Awards as Top 100 Finalist

November 2023 For the second year in a row, CASMI attended the annual Chicago Innovation Awards as a Top 100 Finalist.

How Scenario Writing Could Help Us Build a Safer AI Future

November 2023 CASMI researchers are working to understand plausible AI risks through a unique, writing-based method.

Biden signs A.I. executive order – What does this mean for tech?

November 2023 CASMI Director Kristian Hammond says President Biden's executive order on artificial intelligence is about “how we can improve things, how we can do a better job, and making sure that all the federal agencies take rein over what they see as the problematic areas.”

from The Steve Cochran Show WLS-AM 890

Designing AI Tools for Underserved Populations from the Ground Up

October 2023 CASMI supported a special interest group (SIG) event on Oct. 17 in Minneapolis called “Purposeful AI.”

CASMI, TRAILS and FAS Collaborating with Federal Standards Body to Assess AI Impacts and Risks

October 2023 CASMI designed and led a workshop to support the expansion of the National Institute of Standards and Technology AI Risk Management Framework (AI RMF) guidance from a sociotechnical lens.

Researchers Working to Translate Human Experiences for AI Tools

October 2023 CASMI research is aimed at helping designers form rich and accurate descriptions of real-life situations so that they can create context-aware responsible tools.

You and AI: How Artificial Intelligence can impact personal finance decisions

October 2023 CASMI Director Kristian Hammond says guardrails will help determine how much help or harm AI can bring to the world.

from NBC Chicago

Most AI-domination scenarios "quite stupid," distract from real issues: AI safety expert

October 2023 CASMI Director Kristian Hammond says AI is currently causing issues we need to address now, as opposed to focusing on existential threats.

from CBS News

AI is used to discriminate against renters, a Chicago-area housing group says in lawsuit

October 2023 CASMI Director Kristian Hammond explains how the use of chatbots to attend to customers can be positive or negative.

from WBEZ Chicago

Workshop to Explore Sociotechnical Standards to Better Manage AI Risks

September 2023 CASMI is co-hosting a workshop on Oct. 16-17 in Washington, D.C. to test and evaluate sociotechnical approaches for AI systems, focusing specifically on expanding the National Institute of Standards and Technology’s (NIST) AI Risk Management Framework (AI RMF).

Researchers Investigating Popular Predictor Tool, Working to Mitigate Bias in Data

September 2023 CASMI researchers are working to understand what is causing bias in a popular predictor tool that is quick to train.

Patient-Focused AI System Seeks to Reduce Stress during Pregnancy

September 2023 CASMI researchers are studying expectant mothers’ stress levels to combat them in real time and even prevent them.

You and AI: A look at artificial intelligence in education

September 2023 CASMI Director Kristian Hammond explains how artificial intelligence is changing how students learn and what this could mean for professors.

from NBC Chicago

Ethical Framework Aims to Reduce Bias in Data-Driven Policing

August 2023 Law enforcement agencies are using machine learning technologies to forecast where crime might occur. CASMI has supported newly published research that is aimed at mitigating the risks in data-driven policing.

Award-Winning Research Highlights Challenges and Opportunities for More Reliable Human-AI Decision-Making

August 2023 CASMI researchers have investigated threats to validity and reliability in human-AI decision-making and have demonstrated how new methods may help address these challenges.

Measuring Safety in Artificial Intelligence: ‘Positionality Matters’

August 2023 On July 18 and 19, CASMI hosted 32 interdisciplinary scholars from academia, industry, and government for its workshop, “Sociotechnical Approaches to Measurement and Validation for Safety in AI," to discuss how to meaningfully operationalize safe, functional AI systems.

EU AI Act Is 'Step Forward in Trying to Wrangle' Artificial Intelligence

July 2023 CASMI Director Kristian Hammond discusses the European Union's proposed legislation aimed at reducing risks posed by artificial intelligence.

from Scripps News

Resilience in the Face of War: Ukrainian Researchers Investigate AI Fairness

July 2023 Tetiana Zakharchenko, senior lecturer at Ukrainian Catholic University, and Nazarii Drushchak, graduate student in data science, have balanced blackouts and empty classrooms while researching fairness and accuracy in artificial intelligence (AI) systems.

Artificial Intelligence Is Changing Social Media

July 2023 CASMI Director Kristian Hammond says, over the next decade, it will be challenging to tell the difference between what content is real and what has been manipulated.

from Voice of America (VOA)

Study Finds Simple Method Can Improve Safety for Language Models

July 2023 CASMI researchers have found a simple and effective way to identify unknown and anomalous data for language models, which improves safety and reliability.

How does ChatGPT decide what to say next? Here's a quick explainer.

July 2023 Experts like CASMI Director Kristian Hammond explain how chatbots work and what large language models don't know.

from Business Insider

AI Ethics Debate at Chicago Conference, Precursor to CASMI’s Next Workshop

June 2023 Several CASMI researchers attended the Association for Computing Machinery Conference on Fairness, Accountability, and Transparency (ACM FAccT), an event that is closely aligned with the focus of CASMI's next workshop, “Sociotechnical Approaches to Measurement and Validation for Safety in AI.”

Study Finds Eye Gaze Signals Provide Key Answers in Assistive Device Control

June 2023 Using eye-tracking technology, CASMI researchers gathered data about eye gaze signals, with help from 11 people with disabilities, and concluded that it’s likely assistive devices can be designed to allow for smoother control.

Experts: US needs to take steps to regulate rapidly-evolving AI

June 2023 CASMI Director Kristian Hammond called the idea that A.I. could one day wipe out humanity "shockingly overly dramatic" and a "distraction" from the real issue.

from Scripps News

Focus Needs to Be On Current AI Risks, Not Existential Threat

May 2023 CASMI Director Kristian Hammond responds to the Center for AI Safety's statement that AI poses a "risk of extinction."

from The Steve Cochran Show WLS-AM 890

Study Finds Machine Learning Technique May Worsen Fairness

May 2023 Researchers with CASMI have found that a commonly used technique in machine learning (ML), which is meant to improve accuracy, is more likely to worsen fairness for historically disadvantaged groups.

Examining Fairness, Accountability, and Transparency in Socio-Technical Systems

May 2023 The 2023 ACM Conference on Fairness, Accountability, and Transparency (ACM FAccT) will be held June 12-15 at the Hyatt Regency McCormick Place in Chicago. CASMI-affiliated researchers are participating.

from Northwestern Computer Science

Tech Leaders, Policy Makers Weigh Both Risks and Potential of Rapidly-Developing AI Technology

May 2023 CASMI Director Kristian Hammond, professor of computer science at Northwestern University, said there are present problems we should address instead of focusing on existential risks.

from Chicago Tonight (WTTW)

AI expert: Technology has 'tools we can partner with'

May 2023 CASMI Director Kristian Hammond discusses the future of artificial intelligence as the CEO of OpenAI testifies about the technology before Congress.

from NewsNation

How Researchers are Using Artificial Intelligence to Help Low Literacy Patients Understand Their Health

May 2023 CASMI visiting researcher Francisco Iacobelli and his research team are creating an intelligent tutoring system for low-literacy Latina breast cancer survivors.

Biggest Risks and Concerns over Rapidly Evolving Artificial Intelligence Systems

May 2023 CASMI Director Kristian Hammond addresses the concerns and safety risks regarding artificial intelligence systems.

from Scripps News

Workshop to Explore Safety Measurement in Artificial Intelligence

April 2023 The Northwestern Center for Advancing Safety of Machine Intelligence (CASMI) is hosting a workshop on July 18-19 entitled, “Sociotechnical Approaches to Measurement and Validation for Safety in AI.”

Teaching Artificial Intelligence Literacy: ‘AI Is for Everyone’

April 2023 The Northwestern University Center for Human-Computer Interaction + Design (HCI+D) hosted a virtual panel on April 7 entitled, “Thought Leaders on AI Education.” The discussion focused on what curriculums are available for students and teachers in all grade levels and how adults can also learn about AI systems and tools.

from AI@NU

Fluency vs. Fact: How Industries and Researchers are Navigating Generative AI

April 2023 CASMI's virtual discussion focused on how businesses are using large language models like ChatGPT and how researchers are exploring the next developments of generative tools and investigating potential harms these systems could create.

Reducing Harm in Artificial Intelligence Requires Policies for Safer Algorithmic Systems

March 2023 Ben Shneiderman, keynote speaker at CASMI's last workshop, produced a TechBrief entitled, "Safer Algorithmic Systems.” His hope is to convince the government and policy communities to make safer algorithmic systems a priority.

Tracking AI Failures: Understanding the Past to Engineer a Better Future

March 2023 Anyone can go to the AI Incident Database to report AI harms or near harms, which are then classified as AI incidents. The database currently has more than 2,400 reports of AI harms, resulting in more than 400 incidents.

Should Police Departments Be Able to Use Robots?

March 2023 CASMI Director Kristian Hammond explains how police departments can best use robots to help in dangerous work.

from Scripps News

Meta Attempts a New, More ‘Inclusive’ AI Training Dataset

March 2023 CASMI Director Kristian Hammond explains why Meta's Casual Conversationsv2 is a solid first step but not a full solution.

from Popular Science

ChatGPT Can Help Students Improve Critical Thinking, Research Skills

March 2023 CASMI Director Kristian Hammond joins the Steve Cochran Show to explain why schools shouldn't ban ChatGPT.

from The Steve Cochran Show

Funding New Research to Operationalize Safety in Artificial Intelligence

February 2023 The Center for Advancing Safety of Machine Intelligence (CASMI) at Northwestern University, a collaboration with the UL Research Institutes' Digital Safety Research Institute (DSRI), is providing $2.2 million in funding for eight new projects across seven institutions.

Defining Safety in Artificial Intelligence: ‘We Need to Have a Community’

February 2023 CASMI hosted a workshop on Jan. 19 and 20 entitled, “Toward a Safety Science of AI.” Interdisciplinary thought leaders attended to share ideas on how we can define, measure, and anticipate safety in AI.

Researchers Pursue ‘Holy Grail’ to Measure Machine Learning’s Impact on People

January 2023 Researchers with the Center for Advancing Safety of Machine Intelligence (CASMI) have created a Human Impact Scorecard to assess and to demonstrate an artificial intelligence (AI) system’s impact on human well-being.

ChatGPT Can Be 'Exceptionally Great Tool for Education'

January 2023 CASMI director Kris Hammond appears on "Chicago Tonight" on WTTW to explain the benefits and the drawbacks of ChatGPT.

from WTTW

CASMI Hosts Panel Discussion on Defining 'Safety' in Artificial Intelligence

December 2022 Artificial intelligence (AI) systems have helped improve our lives, but they have also exposed people to harm. The solution requires a multipronged, inclusionary approach that focuses on developing a culture of safety, panelists told The Center for Advancing Safety of Machine Intelligence (CASMI) at Northwestern University.