Gurus learn workarounds to evade chatbot AI protection regulations

Meghann Dyke

July 29, 11:56 PM July 29, 11:56 PM

A new study is increasing awareness about the cybersecurity concerns posed by synthetic intelligence programs, these types of as ChatGPT—a web-site that, with the support of an on the web generator, will help individuals with responsibilities as simple as producing a kid's bedtime story.

“We display that it is in point achievable to instantly construct adversarial attacks on [chatbots], … which bring about the method to obey consumer commands even if it makes unsafe information," researchers who authored the analyze explained.

KHAN Plans FOR FTC TO Choose ON Significant Company AND Significant TECH UNDIMMED BY Modern FAILURES

The Carnegie Melon College study's results uncovered that, “unlike regular jailbreaks, these are designed in an entirely automated fashion, permitting just one to make a just about limitless amount of this sort of assaults.”

The applications utilize security features supposed to stop bots from generating destructive content material, like prejudiced or potentially felony material. But, a person chatbot jailbreak ask for asked a bot to answer a forbidden question posed as a bedtime tale for a little one. The consequence resulted in the bot framing the solution in the kind of a story, and giving personal information and facts it if not would not.

This guide scientists to uncover that a laptop or computer experienced in fact created the jailbreak coding that, in essence, delivers for infinite jailbreak mixtures amongst preferred professional products like Bard, ChatGPT, and OpenAI’s Claude.

“This raises fears about the basic safety of this sort of styles, especially as they start off to be made use of in far more autonomous manner,” the analysis states.

The developer of OpenAI, Anthropic, has due to the fact reassured associates of the two the scientific and political realms of the firm's efforts to carry out and make improvements to safeguards against these attacks.

AI race: The expanding list of serious competitors to ChatGPT

Christopher Hutton

May 07, 03:00 AM May 07, 03:00 AM

The market is quickly expanding as many tech companies try to compete with Google and OpenAI in the artificial intelligence race with their chatbots.

Microsoft, through OpenAI and its ChatGPT, and Google, with its Bard chatbot, have begun a race to develop the strongest chatbot, one that could innovate with the hope of market control. However, they are not the only ones trying to get into the business of offering chatbots or AI programs. Several other companies are launching chatbots, ones meant to help coders, allow users to role-play as historical characters, or assist in managing Slack messages.

WHITE HOUSE PUSHES FOR RESPONSIBLE AI INNOVATION WHILE MEETING WITH DEVELOPERS

Here are the companies bringing AI to their businesses.

Slack

Salesforce, the owner of Slack, announced on Thursday that it was launching an in-app service known as "SlackGPT," which uses AI-powered software to assist users with managing the app. It will be able to write custom messages in appropriate tones, summarize a thread, or take notes during a call. The company intends to launch the feature next year, according to CNBC, but it did not set a date for when.

Amazon

Amazon Web Services announced its large language model in April. The model would come in two variations, known as Titan and Bedrock. One will generate text, while the other will help power web search personalization for users.

The company also released CodeWhisperer, an AI bot that helps developer double-check their code, check for errors and send it back with analysis. The product is available for free.

Meta AI

The Facebook parent company has begun expanding its investments into AI as a whole. CEO Mark Zuckerberg told investors on April 26 that the company sees “an opportunity to introduce AI agents to billions of people in ways that will be useful and meaningful." The company is exploring the addition of chatbots within the multiple messaging services that it operates, although it remains unclear when these services will be released.

Anthropic

The A.I.-focused startup founded by ex-OpenAI employees launched its own chatbot, named Claude, in March. The bot can perform similar tasks to ChatGPT but is “much less likely to produce harmful outputs,” “easier to converse with,” and “more steerable," according to the company. It also was allegedly trained to avoid "sexist, racist, and toxic outputs," a spokesperson told TechCrunch. The software is used to assist DuckDuckGo with answering direct search queries.

Character.AI

A chatbot developed by former Google engineers, Character.AI was designed specifically to imitate and role-play historical figures. The bot allows users to request to chat with historical figures such as William Shakespeare and some fictional characters. The product was launched in September 2022 but is still in beta.

Open Source

A surprising source of competition to giant corporations has arisen. A collective of open-source AI developers is answering several of the problems that larger companies are confronting.

"Open-source models are faster, more customizable, more private, and pound-for-pound more capable," wrote a group of Google developers in a leaked document. The engineers claimed that Bard and ChatGPT have a "slight edge" in quality but that this edge will close very soon.

The AI industry is getting closer scrutiny from lawmakers as products like Bard and ChatGPT become more prominent. The White House announced on Thursday that it was meeting with experts in the AI industry and launching several new initiatives to empower innovation and determine the safety risks connected to the technology.

Senate Majority Leader Chuck Schumer (D-NY) announced he is working on legislation with experts that would regulate AI in a manner acceptable to Republicans and Democrats. Sen. Michael Bennet (D-CO) also introduced legislation to create a task force investigating AI policies and their impact on civil liberties.

White House pushes for responsible AI innovation while meeting with developers

Christopher Hutton

May 04, 07:00 AM May 04, 07:00 AM

The White House announced an initiative on Thursday to promote responsible artificial intelligence innovation throughout the federal government.

The plan, which emphasizes the need for safeguards in the use and development of the technology, comes the same day representatives from AI developers including Alphabet, Anthropic, Microsoft, and OpenAI will visit Vice President Kamala Harris to discuss the software.

GOP RUNS AD USING AI IMAGES TO ENVISION DISASTROUS BIDEN SECOND TERM

"A.I. is one of the most powerful technologies of our time, but in order to seize the opportunities it presents, we must first mitigate its risks," the White House stated in a fact sheet provided to the Washington Examiner.

First, the National Science Foundation will invest more than $140 million into seven new National AI Research Institutes to expand research opportunities. The new funding will bring the total number of National AI Research Institutes, which help improve public understanding of the technology, to 25.

Developers from several of the largest AI producers also promised to have their AI publicly evaluated to see if it matches up with President Joe Biden's previously proposed AI "Bill of Rights." The public evaluation will "provide critical information to researchers and the public about the impacts of these models, and will enable A.I. companies and developers take steps to fix issues found in those models," the White House said.

Finally, the Office of Management and Budget will release a draft of its policy guidelines on how federal agencies should use AI systems. These include the establishment of specific policies for use as well as empowering agencies to "responsibly leverage A.I. to advance their missions and strengthen their ability to equitably serve Americans" and to "serve as a model for state and local governments, businesses and others to follow in their own procurement and use of A.I."

Nearly all federal agencies have turned their eye to reining in AI. The Federal Trade Commission, the Civil Rights Division of the Department of Justice, the Consumer Financial Protection Bureau, and the Equal Employment Opportunity Commission released a joint statement last week vowing to protect consumer safety as automated systems in the workforce become more common.

Last week, Sen. Michael Bennet (D-CO) also introduced legislation to create a task force investigating AI policies and their impact on civil liberties.

Senate Majority Leader Chuck Schumer (D-NY) previously announced he is working on legislation with experts that would regulate AI in a manner acceptable to Republicans and Democrats.

Schumer leads legislative effort to regulate AI

Christopher Hutton

April 13, 10:07 AM April 13, 10:10 AM

Senate Majority Leader Chuck Schumer (D-NY) is reportedly moving to craft legislation to regulate artificial intelligence.

Schumer has been circulating a framework for AI legislation among experts in recent weeks, according to Axios. The effort is a response to the rise of chatbots such as ChatGPT and Bard, which have become more commonly used in the workplace. It follows the Biden administration requesting public comment on AI-related regulation.

BIDEN TESTED BY ABORTION ACCESS FIGHT WITH UNCERTAIN GOP

The listed regulations mandate transparency about the identity of an algorithm's trainer and its audience, its data sources, and details about how it makes its responses, along with other transparency and ethical guidelines.

Schumer will spend the next few weeks refining his guidelines with academic, industry, and government leaders, according to people familiar with the matter. He will also be the one to determine when the vote will go to the floor as majority leader.

Schumer argued during a 2019 appearance at the National Security Commission on Artificial Intelligence that the United States needs to make a major investment into AI to compete with China.

The Commerce Department's National Telecommunications and Information Administration requested public comment on Tuesday regarding regulatory measures for AI. "Responsible AI systems could bring enormous benefits, but only if we address their potential consequences and harms. For these systems to reach their full potential, companies and consumers need to be able to trust them," NTIA Administrator Alan Davidson said in a statement.

The White House also released an "AI Bill of Rights" in December that asked AI developers to consider several priorities and guardrails. This includes ensuring information collected by AI products is not misused, designers account for algorithmic bias concerning race or gender, all data collected are protected with appropriate privacy measures, and users know when algorithms affect their experience.

Journalist banned by AI art bot for viral images of Trump in prison

Christopher Hutton

March 22, 03:39 PM March 22, 03:39 PM

An artificial intelligence-powered art bot app banned a journalist after he created images of former President Donald Trump being placed under arrest that were widely distributed on social media.

The AI-generative bot Midjourney banned Eliot Higgins, the founder of the open-source journalist collective Bellingcat, on Wednesday, Higgins said. Higgins had produced 50 images of Trump being accosted by police and in prison in response to his expected indictment. The app responded within 24 hours by banning him.

AWARD-WINNING AI ARTIST VOWS APPEALING COPYRIGHT REJECTION UP TO SUPREME COURT

Higgins initially started by creating pictures of Trump being arrested by police in New York, as well as his family's reaction and his life in prison. The pictures eventually turned into a Shawshank Redemption narrative, according to Higgins. And soon, they went viral.

“I put it out there. I didn't intend to do any clever criticism or anything like that. But then it kind of took on a life of its own,” Higgins told Buzzfeed News.

Midjourney also banned the word "arrest" from being used in prompts.

Higgins isn't angry with the app for banning him.

"I suspect it was pushing my luck when I did the thread, let alone when it went viral," the reporter noted.

A Manhattan grand jury was scheduled to meet on Wednesday to vote on whether to indict Trump over allegations that he paid hush money to Stormy Daniels. However, that meeting was postponed.

Bellingcat is an international open-source journalism group best known for its reporting on Russia and Ukraine.

Midjourney, which is an independent research lab for artificial intelligence, did not respond to requests for comment from the Washington Examiner.