Is AI biased against certain groups and spreading misinformation and extreme views? | The Brink

Photo by Lidiia Moor/iStock

Artificial IntelligenceBU computer scientists will scrutinize artificial intelligence for socially undesirable behavior with the support of a first-of-its-kind federal AI pilot program.

Millions of people have tried out artificial intelligence tools like ChatGPT, testing their ability to write essays, create art, make movies, improve internet searches, and even explore whether they can provide friendship, companionship, and even romance.

But can and should we trust AI? Is it safe, or does it perpetuate prejudice and spread hatred and misinformation?

Boston University computer scientist Mark Crovela studies a type of AI called large-scale language models. Photo by Mike Spencer.

These are the questions Boston University computer scientist Mark Crovela will explore in a new project backed by a first-of-its-kind National Science Foundation (NSF) and Department of Energy program, the National Artificial Intelligence Research Resource (NAIRR) pilot, which aims to bring a new level of scrutiny to the dangers and promises of AI by providing 35 projects, including Crovela's, with access to advanced supercomputing resources and data.

Clovera, professor of computer science in the School of Humanities and Letters at Boston University and professor and academic chair of the Department of Computing & Data Science, will use high-performance assistance to investigate a type of AI called large language models (LLMs). His goal: to audit LLMs, AI programs trained to explore and summarize data, generate text and speech, and predict socially undesirable behavior. LLMs help power everything from ChatGPT to automated chatbots and smart speaker assistants. Clovera will work on the project alongside Ebi-Maria Terzi, professor of computer science at CAS.

The research resource pilot grew out of President Joe Biden's October 2023 executive order calling for a coordinated federal approach to safely and responsibly govern the development and use of AI, NSF said.

Brink asked Crovella about the rapid expansion of AI, how it is already part of our daily lives, and how the NAIRR Prize can help determine whether AI is trustworthy and safe.

Q&A with Mark Crobella

The Brink: In your project description, you mention that the LLM will soon be widely adopted. How can it impact our lives? Or has it already impacted them?

Crovella: Large-scale language models are rapidly becoming popular with software tools like ChatGPT. LLMs are rapidly gaining adoption in education. Students and teachers alike use them to speed up their work. In social settings, many companies sell LLMs as online companions and assistants. In science, researchers use LLMs to find and summarize key developments in the vast amount of research published every day. Apple, Microsoft, and Meta have all announced that they will be integrating LLMs into their product lines. In fact, ChatGPT has become the fastest-growing new software, reaching 100 million users in just two months, much faster than TikTok or Facebook.

The Brink: What are your goals with this project?

Crovella: Given that millions of people will be interacting with LLM on a daily basis, we think it's important to ask what social values ​​are embedded in these models — whether they contain bias against protected groups, tendencies to promote extreme or hateful views, or conversation patterns that steer users towards unreliable information.

As millions of people will be interacting with LLMs on a daily basis in the near future, we believe it is important to ask what social values ​​are embedded in these models.

Mark Crovella

The question is, how do we evaluate these tendencies in a system as complex as ChatGPT? In our previous work, we studied simpler systems from the outside: we fed them inputs and observed whether their outputs were biased. But when we look at LLMs, this strategy starts to break down. For example, we find that LLMs correctly reject answers to questions about sensitive topics when asked in English, but simply asking the same question in a different language (Portuguese) causes the model to provide answers that it shouldn't have.

Therefore, LLMs cannot be trusted just by looking at them from the outside. The impetus for this new project is to ask whether we can look inside LLMs, observe the representations of concepts they use, observe the logic they follow, and from there detect whether the system might incorporate undesirable social behaviors. In essence, we want to study LLMs in the same way that neuroscientists study the brain with an fMRI machine.

The Brink: Senator Ed Markey (D-MA, 4th District) praised the NAIRR program for accelerating AI research, saying their public purpose-driven projects are crucial at a time when AI development is dominated by big tech companies. Do you agree with this, and why is it important to have public purpose-driven projects in this space?

I take Senator Markey's views to underscore the need for independent research on AI systems. These very impressive systems would not exist without the big tech companies; they have no way of attracting the necessary resources elsewhere. At the same time, many eyes need to be on the question of whether these new systems are helping us. I want to point out that our work relies on the fact that some of the big tech companies are actually open-sourcing their models, providing their code and data for free. This is a socially beneficial act for them, and it's crucial for the kind of work we do.

The Brink: Will this award enable you to do things you couldn't do before?

LLMs are large in two senses: they contain vast amounts of knowledge derived from vast amounts of training data, and they have highly complex systems that use billions of parameters to generate output. As a result, the internal representations used by LLMs are said to be huge and impenetrable. Just processing the sheer amount of information within a modern LLM is a massive computational task. The NAIRR grant gives us access to top national laboratory supercomputing facilities, enabling us to efficiently analyze the inside of a modern LLM.

The Brink: How does your work benefit society and the general public?

Currently, there are over 500,000 different LLMs available to the public. It is certain that in the near future, each of us will have our own personal LLM that knows a lot about us and helps us with various tasks on a minute-by-minute basis. How do we know that these huge and mysterious systems are trustworthy and secure? Our research aims to provide ways to answer these questions.

View related topics:

What Are The Main Benefits Of Comparing Car Insurance Quotes Online

LOS ANGELES, CA / ACCESSWIRE / June 24, 2020, / Compare-autoinsurance.Org has launched a new blog post that presents the main benefits of comparing multiple car insurance quotes. For more info and free online quotes, please visit https://compare-autoinsurance.Org/the-advantages-of-comparing-prices-with-car-insurance-quotes-online/ The modern society has numerous technological advantages. One important advantage is the speed at which information is sent and received. With the help of the internet, the shopping habits of many persons have drastically changed. The car insurance industry hasn't remained untouched by these changes. On the internet, drivers can compare insurance prices and find out which sellers have the best offers. View photos The advantages of comparing online car insurance quotes are the following: Online quotes can be obtained from anywhere and at any time. Unlike physical insurance agencies, websites don't have a specific schedule and they are available at any time. Drivers that have busy working schedules, can compare quotes from anywhere and at any time, even at midnight. Multiple choices. Almost all insurance providers, no matter if they are well-known brands or just local insurers, have an online presence. Online quotes will allow policyholders the chance to discover multiple insurance companies and check their prices. Drivers are no longer required to get quotes from just a few known insurance companies. Also, local and regional insurers can provide lower insurance rates for the same services. Accurate insurance estimates. Online quotes can only be accurate if the customers provide accurate and real info about their car models and driving history. Lying about past driving incidents can make the price estimates to be lower, but when dealing with an insurance company lying to them is useless. Usually, insurance companies will do research about a potential customer before granting him coverage. Online quotes can be sorted easily. Although drivers are recommended to not choose a policy just based on its price, drivers can easily sort quotes by insurance price. Using brokerage websites will allow drivers to get quotes from multiple insurers, thus making the comparison faster and easier. For additional info, money-saving tips, and free car insurance quotes, visit https://compare-autoinsurance.Org/ Compare-autoinsurance.Org is an online provider of life, home, health, and auto insurance quotes. This website is unique because it does not simply stick to one kind of insurance provider, but brings the clients the best deals from many different online insurance carriers. In this way, clients have access to offers from multiple carriers all in one place: this website. On this site, customers have access to quotes for insurance plans from various agencies, such as local or nationwide agencies, brand names insurance companies, etc. "Online quotes can easily help drivers obtain better car insurance deals. All they have to do is to complete an online form with accurate and real info, then compare prices", said Russell Rabichev, Marketing Director of Internet Marketing Company. CONTACT: Company Name: Internet Marketing CompanyPerson for contact Name: Gurgu CPhone Number: (818) 359-3898Email: [email protected]: https://compare-autoinsurance.Org/ SOURCE: Compare-autoinsurance.Org View source version on accesswire.Com:https://www.Accesswire.Com/595055/What-Are-The-Main-Benefits-Of-Comparing-Car-Insurance-Quotes-Online View photos