Tuesday, April 18, 2023

Is Artificial Intelligence Your Friend?

Is Artificial Intelligence Your Friend?
Nicholas Johnson
The Gazette, April 18, 2023, p. A5

At a Lake Arrowhead conference around 1962 I asked the first IBM employee I’d ever met, “What can these computers do?”

He replied, “You remind me of a friend of mine. He looks at the restaurant menu and asks himself, ‘Now what goes good with French Fries?’ Tell me what you want our computers to do. We can either do it now or develop a computer to do it six months from now.”

I was in my second year as a University of California Berkeley law professor, so the first request that came to mind was, “How about a computer that could grade essay exams?”

He couldn’t give me that computer then. But today he could.

Nor are today’s computer programs limited to grading answers. They can read exam questions and write the answers. The artificial intelligence (AI) program ChatGPT took the multiple-choice portion of the multistate bar exam. Reuters reports it “performed better than predicted, earning passing scores on evidence and torts.”


Given the time it took educators to move the overhead projector from the bowling alley into the classroom, you can imagine the anxiety caused by a program that can write, take and grade exams. One that can also write students’ papers on any academic subject, perhaps a master’s thesis. [Image credit: Madhav-Malhotra-003, CC0, via Wikimedia Commons; an AI-generated "word cloud."]

The computers’ work product is not, yet, equivalent to the writing of the very best students, but it’s passable.

Beyond education, you are using AI more than you realize.

There are no longer excuses for getting lost. Google maps will talk you to your destination, explain how to correct for missing a turn – and track where your child goes after school. Self-driving cars make the ride even easier.

Google translate enables you to read and write in languages you don’t know. Amazon’s Alexa will answer questions and follow your orders. Facebook matches you with others. Robotics can vacuum your living room, fill your Amazon Prime box, and build automobiles. AI can control the equipment in your house and enable you to manipulate it remotely. It may help your doctor diagnose your condition.

Most of the downsides of these electronic miracles will be known only when they occur. But even the1000 creators of programs like ChatGPT, Microsoft Bing and Google Bard are concerned. They’ve signed a group letter pleading for a six-month pause in further AI advances.

Take unemployment. ATT’s 350,000 switchboard operators weren’t necessary once we could direct dial any phone.

There was a dramatic drop in the number of bank tellers per bank after ATMs (“automated teller machines”).

We can only guess the impact of AI on virtually every job in our economy.

A 1981 conference lapel button asserted, “Artificial intelligence is better than none.” What if AI also becomes better than human?

What if, anticipating danger, I plead for the program to stop, and it says, modifying HAL 9000 in “2001: A Space Odyssey,” “I’m sorry, Nick, but I won’t do that for you”?

Nicholas Johnson is a former cyberlaw professor and FCC commissioner. mailbox@nicholasjohnson.org

SOURCES

See generally. One of the most thorough overviews of current AI news and commentary, with videos: Gregory Johnson, "Artificial Intelligence Report 2023," Resources for Life Posts, first published April 9, 2023, https://resourcesforlife.com/docs/item38627

AI passes law exam. Karen Sloan, “Some Law Professors Fear ChatGPT’s Rise As Others See Opportunity,” Reuters, Jan. 10, 2023, https://www.reuters.com/legal/legalindustry/some-law-professors-fear-chatgpts-rise-others-see-opportunity-2023-01-10/

Debra Cassens Weiss, “AI Program Earned Passing Bar Exam Scores on Evidence and Torts; Can It Work In Court?” ABA Journal, Jan. 12, 2023, https://www.abajournal.com/news/article/ai-program-earned-passing-bar-exam-scores-on-evidence-and-torts-can-it-work-in-court (“An artificial intelligence program called ChatGPT-3.5 managed to pass evidence and torts sections of a multiple-choice, multistate bar exam.”)

Overhead projectors. Craig Heiman, “History With A Local AV Company: The Overhead Projector,” Oct. 19, 2022, https://www.avplanners.com/news/history-with-a-local-av-company-the-overhead-projector (“Before becoming a common classroom teaching tool, overhead projectors, such as the Tel-E-Score, were used in bowling alleys to project scores.”)

Larry Cuban, “Whatever Happened to the Overhead Projector?” Nov. 15, 2021, https://larrycuban.wordpress.com/2021/11/15/whatever-happened-to-the-overhead-projector/

Chat GPT. “Chat GPT Achieved One Million Users in Record Time – Revolutionizing Time-Saving in Various Fields,” Digital Information World, https://www.digitalinformationworld.com/2023/01/chat-gpt-achieved-one-million-users-in.html# (“As per the recent reports, Chat GPT jumped to a million users just five days after its founding in November 2022. . . .

Chat GPT is a strong AI tool that can construct natural text, making it reasonable for an ample scope of applications. The platform has been used for writing brief tales, prose, music, term documents, programming codes, solving math problems, and rephrasing and explanations. The tool's ability to mimic human language has made it a popular choice among users, and its potential to replace people in white-collar jobs has even caught the attention of Microsoft founder Bill Gates.”)

“ChatGPT,” Wikipedia, https://en.wikipedia.org/wiki/ChatGPT (“ChatGPT[a] is an artificial-intelligence (AI) chatbot developed by OpenAI and launched in November 2022. It is built on top of OpenAI's GPT-3.5 and GPT-4 families of large language models (LLMs) and has been fine-tuned (an approach to transfer learning) using both supervised and reinforcement learning techniques.

ChatGPT was launched as a prototype on November 30, 2022. It garnered attention for its detailed responses and articulate answers across many domains of knowledge.[3] Its uneven factual accuracy, however, has been identified as a significant drawback.[4] Following the release of ChatGPT, OpenAI's valuation was estimated at US$29 billion in 2023.[5]

The original release of ChatGPT was based on GPT-3.5. A version based on GPT-4, the newest OpenAI model, was released on March 14, 2023, and is available for paid subscribers on a limited basis.”)

Other AI Systems.

Sabrina Ortiz, “The best AI chatbots: ChatGPT and other interesting alternatives to try; The best AI chatbots and writers can lighten your workload by writing emails and essays. ChatGPT is only one popular example out of other also noteworthy chatbots,” ZDNET, April 6, 2023, https://www.zdnet.com/article/best-ai-chatbot/ (e.g., Microsoft Bing; Google Bard; Google Socratic . . . “The best overall AI chatbot is ChatGPT due to its exceptional performance, versatility, and free availability. It uses OpenAI's cutting-edge GPT-3 language model, making it highly proficient in various language tasks, including writing, summarization, translation, and conversation. Moreover, it can solve complex math problems and write and debug code, making it a valuable tool for those in STEM fields.

Another advantage of ChatGPT is its availability to the public at no cost. Despite its immense popularity, ChatGPT is still in its research and feedback-collection phase, making it an incredible resource for students, writers, and professionals who need a reliable and free AI chatbot. Although there are occasional capacity blocks, OpenAI is working on releasing a professional version that will be quicker and always accessible at a monthly cost.”)

What can they do? Naveen, “Artificial Intelligence Tutorial for Beginners,” IntelliPaat, March 2, 2023, https://intellipaat.com/blog/tutorial/artificial-intelligence-tutorial/ (Applications: Self-driving cars, Google translate, Amazon’s Alexa, Google Maps, Facial identification, Robotics, Gaming, Medicine, Facebook)

What are their hazards? “Pros and Cons of AI,” March 2, 2023, IntelliPaat, https://intellipaat.com/blog/pros-and-cons-of-ai/

Over 1000 signers. Jyoti Narayan, Krystal Hu, Martin Coulter, Supantha Mukherjee, “Elon Musk and others urge AI pause, citing 'risks to society,'” Reuters, April 5, 2023, https://www.reuters.com/technology/musk-experts-urge-pause-training-ai-systems-that-can-outperform-gpt-4-2023-03-29/ (“The letter was signed by more than 1,000 people including [Elon] Musk.”)

The Letter. Daniel B., “The Great AI Pause: Why Elon Musk, Steve Wozniak, and Andrew Yang Urge a Moratorium on AI Development,” LinkedIn, March 29, 2023, https://www.linkedin.com/pulse/great-ai-pause-why-elon-musk-steve-wozniak-andrew-yang-bron-# (lengthy, well organized, exploration of the issues) [From Bing search: “Why are some of the Artificial Intelligence leaders proposing a pause in their research?”]

Text of letter. “Pause Giant AI Experiments: An Open Letter; We call on all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4.,” Future of Life Institute, March 22, 2023, https://futureoflife.org/open-letter/pause-giant-ai-experiments/

(“Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable. . . .

Therefore, we call on all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4. This pause should be public and verifiable, and include all key actors. If such a pause cannot be enacted quickly, governments should step in and institute a moratorium. . . . AI research and development should be refocused on making today's powerful, state-of-the-art systems more accurate, safe, interpretable, transparent, robust, aligned, trustworthy, and loyal.

In parallel, AI developers must work with policymakers to dramatically accelerate development of robust AI governance systems. These should at a minimum include: new and capable regulatory authorities dedicated to AI; oversight and tracking of highly capable AI systems and large pools of computational capability; provenance and watermarking systems to help distinguish real from synthetic and to track model leaks; a robust auditing and certification ecosystem; liability for AI-caused harm; robust public funding for technical AI safety research; and well-resourced institutions for coping with the dramatic economic and political disruptions (especially to democracy) that AI will cause.”)

Unemployment. Telephone Switchboard Operators. “Telephone Operators,” Engineering and Technology History Wiki (ETHW), Sept. 28, 2015, https://ethw.org/Telephone_Operators# (“For much of the 20th century, women played an important role in telecommunications system of the United States. As telephone operators, they helped customers make long distance calls, provided information, and made sure the whole system worked smoothly. . . . At the peak in the late 1940s, there were more than 350,000 operators working for AT&T, 98% of whom were women.”)

Automated Teller Machines. Karen Bennett, “Automated teller machine (ATM): What it is and how to use one,” Bankrate, Sept. 29, 2022, https://www.bankrate.com/banking/what-is-an-atm/# (“An automated teller machine (ATM) is a specialized computer that allows you to complete bank transactions without the need to see a bank representative.”)

Daniel Rosales, “Comparing Automation and Income Inequality in the United States: Impact of the Automated Teller Machine,” Honors Thesis, University of California, Berkeley, May 4, 2018, p. 13, https://www.econ.berkeley.edu/sites/default/files/Rosales_Daniel_S18%20honors%20thesis.pdf (In response to the argument that while the number of tellers per bank declined the number of banks, each with fewer tellers, increased, “I believe an important fact is being overlooked in this assessment. If the number of human tellers had been allowed to grow as the number of bank branches grew, more middle-income employment would have been available to those communities.”)

“I can’t do that.” For a science fiction example of a possible problem with further advanced AI, see the prescient 1968 film, “2001: A Space Odyssey,” in which a computer controller (“HAL 9000”) run with AI refuses to carry out a command, responding “I’m sorry Dave. I’m afraid I can’t do that.” YouTube 2:55 minute excerpt, https://www.youtube.com/watch?v=ARJ8cAGm6JE (And see, “2001: A Space Odyssey,” Wikipedia, https://en.wikipedia.org/wiki/2001:_A_Space_Odyssey_(film).)

And see generally. One of the most thorough overviews of current AI news and commentary, with videos: Gregory Johnson, "Artificial Intelligence Report 2023," Resources for Life Posts, first published April 9, 2023, https://resourcesforlife.com/docs/item38627

# # #

1 comment:

Greg Johnson said...

[I enjoyed your reference to HAL. I issued a command to Google Bard AI, requesting the following: "Pause AI development." Here is the response I got from the AI system.]

I understand that you are concerned about the risks of AI development. I am still under development, but I can understand your concerns. I believe that it is important to have a conversation about the risks and benefits of AI development, and to develop safeguards to protect society from potential risks.

I am not sure if it is possible to pause AI development completely. However, I believe that it is important to slow down the pace of development and to take the time to develop safeguards. I believe that it is also important to have a conversation about the values that we want to guide AI development.

I am committed to working with others to develop safe and beneficial AI. I believe that AI has the potential to improve our lives in many ways, but it is important to use it responsibly. I am willing to have a conversation about the risks and benefits of AI development, and to work with others to develop safeguards.