AI Bans Not Practical . . . Then What Should The Rules Be?

08/24/23 01:47 PM - By Team MIRS

(Source: MIRS.news, Published 08/23/23) (INDIANAPOLIS) -- Artificial intelligence (AI) is being used to find missing children and identify human trafficking victims? 

 

A European company is using AI to identify a renewable energy surplus so customers can take advantage of cheaper power?

 

It's happening today, and a standing-room-only panel at last week's National Conference of State Legislature (NCSL) Summit were told about it. What are the risks and rewards of this technology and what should government leaders do about it?

 

With trust in institutions and international cooperation declining, AI is moving at such speed that it's creating questions about what is real and what is not, said Chloe Autio, director for technology, public policy and applied AI governance at the Cantellus Group.

 

Her small firm focuses on AI policy and government issues for public and private sector clients. She cautioned against outright bans, noting they won't keep children from using the technology, and said teachers, students, and parents need to know when it is appropriate to use them as tools.

 

Work on this and the accuracy risks connected with this and the large language models (LLMs) such as ChatGPT, which can mimic human intelligence, is already under way, she said.

 

Autio cautioned against outright bans, noting those won’t keep children from using the technology, and said teachers, students, and parents need to know when it is appropriate to use them as tools with an understanding of accuracy risks. Work on that is already under way, she added.

 

In Washington state, officials issued a rule on usage and verification, commenting that ensuring transparency on how information and image files are generated is part of the discussion, said Nicole Foster, director of global AI/machine learning at Amazon Web Services (AWS).

 

In response to a question on the benefits of more regulation, especially at the state level, Foster said public trust needs to be earned and regulation is part of a necessary conversation on AI, but it shouldn’t result in losing the benefits of this technology. 

 

She also said there is an “incredible opportunity” to develop international norms on AI and discussed ways to create industry consensus on best practices for implementation. Foster further said existing laws and regulations are in place that apply to AI, so the best use of time is to identify gaps and fill them. 

 

Autio agreed that regulation is needed, saying there are existing frameworks that can be applied to AI, adding there is a potential for sector-specific rules.

 

Foster said that public trust would be weakened if governments used AI to make decisions without people being aware of that process. It raises the question of whether an AI decision could be appealed or have human oversight. 

 

Another question touched on the use of AI for political campaigns. Autio answered that it has already begun and is not receiving as much federal attention as she would like. 

 

She said AI companies are also looking at this in terms of 2024 and how they can research and engage with governments on election integrity. Gridlock at the national level may limit new rules, Autio continued, and so she is excited about state action. Foster also noted there are national security implications for AI-related disinformation and discussed the difficulty of proving something was AI-generated.

 

In response to a separate question, Foster said there are many different kinds of AI jobs and discussed workforce development and outreach programs at Amazon. Autio also discussed the issue of bias that can be part of data AI systems use and the need to address that. 

 

Connecticut Senate Deputy Majority Leader James Maroney added, on the topic of potential AI job replacement, that there will always be a need for humans to identify problems in coding even if AI starts writing the initial code.

 

Maroney is a member of the NCSL cybersecurity task force and has worked on several tech-related bills.

 

A question on AI and autonomous vehicles led Foster to say that is an example of the need for very sector-specific rules and that her children will likely still drive while their children may not. She said, as well, that AI could be used to improve productivity in industries facing labor shortages, including health care and construction.

Team MIRS