What policymakers need to know about artificial intelligence
Artificial Intelligence 101: Understanding The Power Of AI And Its Dangers
PHILADELPHIA (WPVI) -- Artificial intelligence is rapidly evolving.
We hear about it often, but many people are uninformed about how they're currently using it and don't know how to incorporate it in their daily lives.
AI refers to the simulation of human intelligence in computers or machines that are programmed to think and learn like humans.
The technology is designed to perform tasks like visual perception, speech recognition, decision making and language translation.
From Apple's Siri to Google's Alexa and many customer service chatbots, you're likely already enmeshed in AI.
Social media networks even use the technology to analyze your behavior and moderate hate speech.
Inside the University of Pennsylvania's School of Engineering and Applied Science, professor Chris Callison-Burch showed us ChatGPT, which is a system that allows people to create images with a simple text prompt.
We input "robot sitting at desk working." Within seconds images appeared.
"That's a pretty good image," said Chad Pradelli.
"It's pretty good," replied Callison-Burch.
Professor Callison-Burch also scripted a movie in ChatGPT. Fantastical but life-like characters came to life through its image generator.
In everyday life, ChatGPT and similar programs can also be used to help plan a presentation, write a resume, or even brainstorm ideas.
"So, I am typing, 'Let's brainstorm ideas of an interview about AI introducing tools to the general public,'" said Callison-Burch.
Within a few seconds, ChatGPT presented us with detailed topics in an organized format to discuss.
"Took longer for me to type in a prompt than it did for GPT to generate a coherent output," he said.
Professor Callison-Burch said everyone should be experimenting with AI technologies.
Some are free, while others require subscriptions or a combination of both.
"Ultimately it's going to become part of everyone's everyday job," said Callison-Burch.
"Should people be concerned about their careers and jobs?" asked Chad Pradelli.
"I think it's still too early to see how widespread this effect will be," Callison-Burch replied.
But users of generative AI need to be careful. Traditional AI excels at analyzing data and performing specific tasks. Generative AI goes a step further, focusing on creating new content like text, images, music and more. It is also designed to learn and make decisions.
"I personally would not be comfortable turning over to generative AI model, access to my bank accounts, and having it act on what it perceives my instructions to be," said professor Michael Kearns.
Kearns is also a professor of AI at the University of Pennsylvania. He and others said AI can "hallucinate," meaning output things that are confidently stated but not grounded in fact -- meaning simply not true.
It's also created a flood of what's called "deepfakes," which are artificial images or video manipulated to influence elections, harm individuals, or just spread disinformation.
"I think there's going be some serious policy issues and societal issues that we're all going to have to grapple with," said Kearns.
Other AI concerns deal with bias and discrimination, and concerns generative AI will understand it is smarter than humans and best to make decisions.
Pennsylvania and several other states have begun passing legislation in an effort to control the wave of artificial intelligence coming our way.
Copyright © 2024 WPVI-TV. All Rights Reserved.
Artificial Intelligence
Simply put: No field in computing is as hot or heavily hyped right now as artificial intelligence. Once a science fiction dream, A.I. Is now a big part of our everyday lives; whether it's commonplace tech like fraud prevention systems, online recommendations, or the speech recognition tools driving smart assistants — all the way through to innovative new technologies like self-driving cars and autonomous delivery robots. Want to know why A.I. Matters? In need of a primer on the difference between deep learning and reinforcement learning? Scared at the possibility of machine superintelligence? Helpfully, we've rounded up all our news, guides, and assorted other A.I. Coverage in one central location.
Artificial Intelligence At ETH Zurich
Artificial intelligence (AI) refers to technology that enables computers to help humans with tasks that require intelligence to solve. (Illustration: ETH Zurich / John Devolle)
The applications and methods of artificial intelligence (AI) are becoming increasingly visible and present in the everyday life of science, business and society. AI and machine learning not only affect private users and industrial processes, but are also changing the way in which researchers and computers share their work (see box below).
In principle, AI can help to expand methods in every field of research, and both AI and machine learning are now firmly established in teaching, research and knowledge transfer at ETH Zurich.
Innovation through connecting research areasETH Zurich's strength in AI lies in:
A combination of excellence in the general aspects of AI and cutting-edge research in the individual disciplines offers huge potential for innovative AI methods that are reliable, explicable and trustworthy.
The ETH AI Center – a central hub for AIBuilding on its existing strengths, ETH Zurich opened the ETH AI Center in October 2020. The new ETH AI Center will lead the way towards trustworthy, accessible, and inclusive AI systems for the benefit of society.
It unites researchers of AI foundations, applications, and implications across all departments at ETH. Starting with the involvement of 29 professorships, its own premises and new Fellowship programmes, the centre will reinforce ETH's strong position in research into this key technology.
The ETH AI Center will join forces with the best AI research institutes in Europe and beyond to accelerate progress, support start-ups and collaboration in industry, and promote the next generation of AI researchers, managers and entrepreneurs. The ETH AI Center is part of the European AI network external pageELLIScall_made and a platform for dialogue between science, business, politics and society.
The ETH AI CenterAs ETH Zurich's central hub for artificial intelligence, the ETH AI Center brings together researchers of AI foundations, applications, and implications across all departments.
Visit the centre's website
Rising student numbers in AI
The student numbers at ETH reflect the increased importance of AI: in 2012/13, just a few hundred students attended a course in machine learning – this figure has now risen to almost 4,000. "Introduction to Machine Learning" is the most popular lecture. Every ETH department has students who attend courses in AI.
Since 2017, ETH has been responding to this demand with an additional Master's programme in data science and a diploma of advanced studies in data science.
Wide range of ETH spin-offs in the AI field
ETH spin-offs from the ICT (information and communication technology) sector generally account for a high proportion of newly established ETH spin-offs, and this share has recently increased further still.
In the last three to five years, the number of new start-ups and ETH spin-offs with an AI focus has also increased. The founders include students, graduates and professors.
ETH spin-offs that use AI methods are active in a wide range of areas, as illustrated by the following selection: real estate (UrbanDataLab), pharma (aiNET, deepCDR), cybersecurity (Exeon Analytics, Xorlab, Futurae), model development (Modulos, LatticeFlow), sewer inspections (Hades Technologies), autonomous robots and drones (Sevensense, Voliro, SeerVision).
An overview of further ETH spin-offs in AI and other areas is available from ETH transfer. The ETH AI Center contributes to strengthening AI start-ups and entrepreneurship.
Artificial intelligence and machine learningArtificial intelligence (AI) refers to technology that enables computers to help humans with tasks that require intelligence to solve.
One important area of artificial intelligence is "machine learning", which has its roots in statistical and data-driven processes. In machine learning, a computer uses training data to learn independently how to identify patterns and regularities in datasets.
Such processes can generate valuable results, particularly in the case of very large, complex or heterogeneous data sets. AI methods complement the researchers' creativity and often deliver surprising suggestions that researchers have not considered.
Artificial Intelligence in the Globe
"Everything AI?" The rise of human-machine partnerships. Globe – the magazine of ETH Zurich and ETH Zurich Alumni 3/2020
International cooperations in AIResearch in AI (selection)
Machine learning and artificial intelligence in the ETH-News
Comments
Post a Comment