Earlier this year, new york times Columnist Kevin Roos had a two-hour conversation with Microsoft’s popular new chatbot, Bing Chat. The “conversation” started innocently enough, but before it ended, the chatbot proclaimed its love for Ruth, saying: You are not happy because you are not in love. You are not in love because you are not with me. ”
When Ruth detailed his encounters with chatbots in his column, the story went viral and led to extensive testing on the part of users to see if they could get similarly spooky responses from chatbots.
Other journalists point to another common problem with generative AI, which if left unchecked can cause widespread problems for society at large. Reporters asked ChatGPT a simple question: new york times What was the first report on “artificial intelligence”? ”
The chatbot said, “It was an article about a seminal conference at Dartmouth College on July 10, 1956, titled ‘Machines will learn and be able to solve problems, scientists predict’. ‘ said.
but times He pointed out that the article was not real, even if the conference purportedly covered in the article actually happened. “ChatGPT is just a hoax,” the newspaper said. “ChatGPT not only gets things wrong sometimes, it can also fabricate information. Names and dates. Medical explanations. Book plots. Internet addresses. Even historical events that never happened.”
Generative AI, enabled by tools like OpenAI’s ChatGPT, Bing Chat, Google’s Bard, and Anthropic’s Claude, is at the forefront of the AI debate in 2023. These systems can generate text, images, and other media by learning from the patterns and structure of their inputs. Train data and create new data to perform various tasks. They have applications in writing, producing works of art, software development, administrative support within organizations, and more.
As these tools become more prevalent, their impact will become even more apparent and urgent in the coming months.
Until now, most people have seen generative AI as a novelty or a form of entertainment. What they don’t know is that AI is already causing a lot of harm. They may not understand the scale of the problem, so they don’t see it. It’s bigger than most people can comprehend.
Generative AI is rapidly becoming part of millions of workflows, with potentially profound implications for democracy, public health, economies, and infrastructure.
Over the next two years, generative AI will become part of Alexa, Siri and Google Maps navigation systems. It will be in our refrigerators, bedrooms, classrooms, offices, and cars. Generative AI will appear everywhere, every minute, every second. When we speak, it listens, draws conclusions, and makes decisions.
If generative AI processes are unchecked, all the systems we rely on in our daily lives are at risk.
Because AI has to learn in the real world, the data we’re collecting today in public health could be misused a few years from now and have devastating consequences.
Consider our collective experience with COVID-19. We relied on public health authorities, traditional media and public health websites. We looked to social media, but most of the time there were real people acting as the sources of the information we found, meaning that the credibility and reputation of that information depended on the quality of that information. I knew you were doing
Generative AI tools can easily be used to unintentionally create inaccurate information—misinformation in the form of disinformation that is intentionally disseminated to deceive. This may be transmitted in the form of text, images or videos and may be transmitted on a large scale. AI will be able to create sophisticated deepfakes that are indistinguishable from the real thing.
And because such misinformation or disinformation is generated in individual sessions and does not appear on publicly accessible web pages, it is difficult to detect. They quietly win hearts and often undermine them. Human moderation would be nearly impossible.
Pandemics are tragic, but unregulated generative AI could make the next one even worse.
There is no going back. Generative AI takes hold. The challenge for us is to make sure that it works for society, not hurts it. Generative AI helps improve productivity, everyday life, health, work, and the economy. But we must now recognize that we are giving this powerful new technology access to a very fragile ecosystem. To protect that ecosystem, the scientific community has a duty to speak up. Please voice your objections or concerns.
With generative AI making such a big presence in society, where does regulation begin?
Trying to regulate users is impractical, so the focus should be on cracking down on AI developers. This is where most problems can occur.
Continuous stress testing at the source is required to ensure that generative AI applications are not only functioning properly, but in an ethical manner.
Regulatory mechanisms may need to be established at the federal and state level. Government, industry and the research community on university campuses need to collectively commit to long-term cooperation, transparency and accountability at the highest levels.
ChatGPT and other generative AI tools may be far from putting us on the brink of extinction, but the risk is real and immediate. The future of our world and future generations is at stake.
Tinglong Dai is Professor of Operations Management and Business Analytics at Johns Hopkins University Carey School of Business.
Copyright 2023 Nexstar Media Inc. All rights reserved. You may not publish, broadcast, rewrite or redistribute this material.
