Senate ramps up AI efforts, but it’s taking time
Senate leaders are planning hearings and briefings on artificial intelligence to consider possible next steps in tackling the rapid development of tools like ChatGPT.
But despite lawmakers ramping up their efforts, there are few calls for immediate action, and despite Silicon Valley’s rapid advance in AI, major companies are preaching patience. ing.
Senators will hold hearings on Wednesday to discuss how generative AI tools like ChatGPT that can create conversational responses to human questions could shake intellectual property, patent and copyright laws. I examined whether there was a possibility. It was the first in a series of follow-up sessions scheduled following last month’s high-profile hearing with the OpenAI CEO. Sam Altman.
Lawmakers heard testimony from academics and representatives of Google and pharmaceutical giant Novartis.
“I think this is one of the countless hearings we need to get right,” the senator said. Tom Tillis The Republican head of the Senate Judiciary Committee’s Intellectual Property Subcommittee (North Carolina) spoke about the role of AI and its commitment to IP.
Sen. Richard Blumenthal The (Democratic, Connecticut) committee, which invited OpenAI chiefs last month, said the two sessions were “just the beginning” of the effort.
Another Senate committee headed by a Senator. John Ossoff (Democrat of Georgia) announces it will hold its own hearing next week to “advance Senate research on machine learning and AI” with a focus on “human rights impacts of transformative artificial intelligence technologies” bottom.
Senate Majority Leader Charles E. Schumer (New York) said: Development of “upper framework” for AIwhich this week announced three separate senator-only briefings on AI this summer, including the House’s first closed session on the subject.
“The Senate must develop its expertise on this pressing issue,” Schumer and other senators wrote to colleagues on Tuesday.
Sen. Mike Roundes Mr. (RS.D.), who co-signed the letter with Mr. Schumer, told my colleague: Kat Zakrushevsky On Wednesday, Congress announced it would hold a briefing before it goes into recess in August. But he did not specify a target date for the announcement of the AI bill.
“We want to do it right,” said Lowndes. “Let’s do no harm in this process.”
Rounds said Schumer’s group has yet to decide whether AI-related legislation will come in a single package or be split among various committees. He raised the idea that the AI Act could work like the Defense Authorization Act.
“We’re going to bring all these ideas and start blending this bill very quietly behind the scenes in a way,” he said. But after that, the bill will be open to public discussion and debate, and other members can submit amendments.
Lowndes also suggested lawmakers could create a task force focused on AI, with Rep. Michael Waltz (R-Florida) putting forward the idea. floating recently.
Rounds said lawmakers were worried because the US hostility “wouldn’t stop.”
“We recognize that time is of the essence and there is real value in getting this going. , I want a really high quality product,” Lowndes told Cat.
Taken together, the remarks suggest that while there is growing interest and concern about AI in the Capitol, legislation is unlikely to be imminent.
Meta launches task force to combat child abuse content on Instagram
Our colleague said Meta has set up a task force to investigate how child abuse material is being disseminated and sold on the Instagram platform. Naomi Nicks report.
“A new effort by Facebook’s parent company has found a large network of accounts believed to be operated by minors openly promoting self-created child sexual abuse content for sale. It follows a report by the Stanford Internet Observatory that did this,” Naomi wrote.
Buyers and sellers of the material were able to connect through Instagram’s direct messaging feature and leverage recommendation algorithms to create more effective ads from the material, the report found. Sales network sizes ranged from 500 to 1,000 accounts at times.
The company said it would dismantle 27 abusive networks individually from 2020 to 2022 and deactivated more than 490,000 accounts in January for violating child safety policies.
“The findings give us more insight into how internet companies have struggled for years to find and prevent sexually explicit images that violate their own rules from circulating on social networks. It provided insight,” Naomi’s report added. In 2021, the company paused development of its Instagram app for children after backlash from some policymakers and child welfare advocates.
FBI warns of AI tools being used for sexual extortion and harassment
The FBI has warned that AI tools are being used to create sexually explicit material for intimidation and blackmail. Rafael Sutter Reuters reported.
Sutter writes: “In a warning circulated this week, the agency recently saw a rise in extortion victims who claimed they were targeted using doctored versions of harmless images taken from online posts, private messages, or video chats. He said he observed that
The agency said the images appeared “real” and in some cases were used against children. “Once widespread, victims may face significant difficulties in preventing the continued sharing or removal of manipulated content from the internet,” the warning reads.
The FBI did not provide details on what AI imaging tools were used in the act, nor did it respond to requests from Reuters for additional information. The agency said advances in technology “continuously improve the quality, customizability and accessibility of AI-powered content creation.”
AI-generated images have been a flashpoint for artists and photographers concerned about copyright infringement, but such content has emerged in a more alarming context that risks causing social unrest. I’m here.
Dutch privacy regulator wants more details about how ChatGPT uses personal data
The Dutch privacy watchdog asked ChatGPT owner OpenAI on Wednesday to explain how its AI chat tool collects personal data when training models to behave. Sarah Jacob Bloomberg News reported.
“In a letter to Microsoft-backed OpenAI, the watchdog asked whether questions typed into ChatGPT would be used to train its algorithms. We also researched how to use it,” writes Jacob.
“The data may include sensitive and highly personal information, for example, when someone asks for advice about a marital dispute or a medical issue,” the regulator said in a statement.
The survey comes as global policymakers launch frameworks to crack down on AI technology amid the booming popularity of ChatGPT, Google’s Bard and Microsoft’s Bing AI chatbots.
EU regulators are set to vote next week on groundbreaking legislation assigning risk categories to various AI technologies, while US lawmakers are gearing up for briefings on the subject, laying the groundwork for AI regulatory plans. there is
Senator Hawley Joins AI Chat (Axios)
Trudeau says Google, Meta are using ‘bullying tactics’ against Canada news bill (Reuters)
Microsoft to bring OpenAI’s GPT model to government cloud customers (Reuters)
Pornhub asks users and big tech companies to help as states introduce age verification laws (CNN)
FOX News says Tucker Carlson breached contract (Axios)
German regulator says Meta plans new profile for Facebook, Instagram users (Reuters)
Companies want to use AI tracking to improve your job performance (Daniel Abril)
Best Air Quality Index app to check nearby wildfire smoke (Tatum Hunter)
Facebook whistleblower Francis Haugen warns of algorithm dangers (Bethany McLean)
TikTok accused by bipartisan senators of ‘inaccurate claims’ to Congress (Forbes)
- General Surgeon Vivek H. Mercy Testifying before Senate Help Committee on Youth Mental Health at 10am
- representative Robin L. Kelly (Illinois) and Jay Overnolty (Republican, Calif.) 11:00 a.m. Speaking at the Center for Bipartisan Policy on the Emerging Tech Workforce event
- Federal Communications Commission to convene June public meeting at 10:30 a.m.
- NTIA Algorithmic Justice Senior Advisor Ellen Goodman Speaking at the German Marshall Foundation Event on AI Governance and Accountability from 12pm
- Cato Institute hosts biometrics and privacy event at 1pm
or‘That’s it for today — thank you for joining us! Tell others to subscribe of technology 202 here. For tips, feedback or greetings, twitter again e-mail.
