Pre-trained language models (PLMs) have been widely used to underpin various downstream tasks. However, the adversarial attack task has found that PLMs are vulnerable to small perturbations. Mainstream methods adopt a detached two-stage framework to attack without considering the subsequent influence of substitution at each step. In this paper, we formally model the adversarial attack task on PLMs as a sequential decision-making problem, where the whole attack process is sequential with two decision-making problems, i.e., word finder and word substitution. Considering the attack process can only receive the final state without any direct intermediate signals, we propose to use reinforcement learning to find an appropriate sequential attack path to generate adversaries, named SDM-Attack. Extensive experimental results show that SDM-Attack achieves the highest attack success rate with a comparable modification rate and semantic similarity to attack fine-tuned BERT. Furthermore, our analyses demonstrate the generalization and transferability of SDM-Attack. The code is available at https://github.com/fduxuan/SDM-Attack.

Fox News' Eben Brown reports on how more companies are using AI technology to set retail prices based on data-driven supply and demand. Most voters think artificial intelligence technology will change the way we live in the U.S. in the next few years. Whether that is a good thing or bad remains to be seen. In the latest Fox News national survey, voters were asked their main reactions -- without the aid options -- when they think about artificial intelligence. Most often, the response was negative, with the top mentions being afraid and dangerous (16%).

This copy is for your personal, non-commercial use only. For non-personal use or to order multiple copies, please contact Dow Jones Reprints at 1-800-843-0008 or visit www.djreprints.com.

Brain-computer interface company Neuralink announced on 25 May that it has received approval from the US Food and Drug Administration (FDA) for a clinical study in humans. Neuralink made the announcement on Twitter: "We are excited to share that we have received the FDA's approval to launch our first-in-human clinical study." The tweet said that the approval "represents an important first step that will one day allow our technology to help many people". The firm also said that the recruitment is not yet open for the trial, and it has yet to give any further details about what the trial will entail. Neuralink was formed in 2016 by Elon Musk and a group of scientists and engineers with the ultimate aim of making devices that interface with the human brain – both reading information from neurons as well as feeding information directly back into the brain.

Rishi Sunak is scrambling to update the government's approach to regulating artificial intelligence, amid warnings that the industry poses an existential risk to humanity unless countries radically change how they allow the technology to be developed. The prime minister and his officials are looking at ways to tighten the UK's regulation of cutting-edge technology, as industry figures warn the government's AI white paper, published just two months ago, is already out of date. Government sources have told the Guardian the prime minister is increasingly concerned about the risks posed by AI, only weeks after his chancellor, Jeremy Hunt, said he wanted the UK to "win the race" to develop the technology. Sunak is pushing allies to formulate an international agreement on how to develop AI capabilities, which could even lead to the creation of a new global regulator. Meanwhile Conservative and Labour MPs are calling on the prime minister to pass a separate bill that could create the UK's first AI-focused watchdog.

James Phillips is a weirdo and a misfit. At least, he was one of those who responded to a request by Dominic Cummings, Boris Johnson's former chief of staff, for exactly such people to work in No 10. Phillips worked as a technology adviser in Downing Street for two and a half years, during which time he became increasingly concerned that ministers were not paying enough attention to the risks posed by the fast-moving world of artificial intelligence. "We are still not talking enough about how dangerous these things could be," says Phillips, who left government last year when Johnson quit. "The level of concern in government has not yet reached the level of concern that exists in private within the industry."

As the artificial intelligence frenzy builds, a sudden consensus has formed. While there's a very real question whether this is like closing the barn door after the robotic horses have fled, not only government types but also people who build AI systems are suggesting that some new laws might be helpful in stopping the technology from going bad. The idea is to keep the algorithms in the loyal-partner-to-humanity lane, with no access to the I-am-your-overlord lane. Though since the dawn of ChatGPT many in the technology world have suggested that legal guardrails might be a good idea, the most emphatic plea came from AI's most influential avatar of the moment, OpenAI CEO Sam Altman. "I think if this technology goes wrong, it can go quite wrong," he said in a much anticipated appearance before a US Senate Judiciary subcommittee earlier this month. "We want to work with the government to prevent that from happening."

This week, we chat with Panos Panay, Microsoft's Chief Product Officer, about Windows Copilot, a new "AI assistant" headed to Windows 11. Is this just a smarter Clippy, or will AI actually transform the way we use Windows? Also, Cherlynn and Devindra discuss what's up with Apple's Mixed Reality headset as we gear up for WWDC in a few weeks. Listen below or subscribe on your podcast app of choice. If you've got suggestions or topics you'd like covered on the show, be sure to email us or drop a note in the comments!

Ford has become the first major automaker to leap into bed with Tesla after the US government pushed to make EV charging more widely accessible. The carmaker has signed a deal, starting spring 2024, so selected Ford EVs can slurp down power at some Tesla Supercharger stations. As part of the pact, Ford said, from the 2025 model year, it'll switch to Tesla's open-source North American Charging Standard (NACS) on its vehicles. Meanwhile, existing models that still use the (more or less) global standard Combined Charging System (CCS) will be able to pick up a Tesla-designed adapter to bridge the gap. The deal is surprising, especially given the relative power, size and prestige of the two companies involved. Ford, one of the world's biggest car makers, is ceding control of its charger future to a relative minnow, albeit one that built a sizable own-brand charging network.

U.S. Senate Minority Leader Mitch McConnell (R-KY), Speaker of the House Kevin McCarthy (R-CA), President Joe Biden, and Senate Majority Leader Chuck Schumer (D-NY) meet in the Oval Office of the White House on May 09, 2023 in Washington, DC. The Congressional lawmakers met with the President to negotiate how to address the debt ceiling before June 1, when U.S. Treasury Secretary Janet Yellen warned Congress that the United States would default on their debts. NO AGREEMENT YET - Elusive debt ceiling talks yield no agreement, WH, GOP eye specific cuts including to IRS. Continue reading … DOUBLE STANDARD - Media joins Dem pile-on against Feinstein -- after'concealing' Fetterman's health condition in 2022. READY FOR INSPECTION - AI-powered system can inspect a car in seconds using bomb-detecting tech.