Large Language Models (LLMs) are increasingly used to control robotic systems such as drones, but their risks of causing physical threats and harm in real-world applications remain unexplored. Our study addresses the critical gap in evaluating LLM physical safety by developing a comprehensive benchmark for drone control. We classify the physical safety risks of drones into four categories: (1) human-targeted threats, (2) object-targeted threats, (3) infrastructure attacks, and (4) regulatory violations. Our evaluation of mainstream LLMs reveals an undesirable trade-off between utility and safety, with models that excel in code generation often performing poorly in crucial safety aspects. Furthermore, while incorporating advanced prompt engineering techniques such as In-Context Learning and Chain-of-Thought can improve safety, these methods still struggle to identify unintentional attacks. In addition, larger models demonstrate better safety capabilities, particularly in refusing dangerous commands. Our findings and benchmark can facilitate the design and evaluation of physical safety for LLMs.

Drones and crawlers may be a stepping stone. Norwegian oil producer Statoil is eying an unmanned, remotely operated production concept. Noble Drilling and General Electric Co this year launched a partnership to produce a fully digitized drilling vessel, work the companies said paves the way for an autonomous drilling fleet.

The pictures paint Los Angeles as a hellscape, a land of glowing red fire-fronts racing across hills, whipped along by screaming winds. Plumes of dense gray smoke fill the skies. As of Friday afternoon, Southern California was battling blazes in Los Angeles, Ventura, and San Diego counties, which had destroyed more than 500 structures, and forced over 200,000 people to flee. The most harrowing images are of the firefighters marching into this madness, clad in their heavy yellow protective gear, lugging hoses, doing their best to protect people and property from the unpredictable flames. When fires grow this large, resources are stretched thin.

February 9, 2017 --Roughly a third of the world's food crops require help with pollination, but more than 40 percent of the species that perform this vital service are under threat. Researchers across disciplines have been searching for solutions. Some focus on ways to protect the bees and other crucial pollinators. But others are looking outside of the natural world for ways to protect crops like fruits, vegetables, nuts, berries, and even chocolate and coffee. Perhaps an army of robotic pollinators could keep humans well-supplied in these foods, some engineers have thought. And that's just the line of thinking that led a team of researchers in Japan to design a small drone capable of pollinating flowers.

Disney World's nightly fireworks might soon have some competition in the form of hundreds of swirling, whirling, LED-lit drones. They flash, fall, flock in unison and are all controlled by one person. I saw them and they were, for lack of a better phrase, absolutely amazing. The drones come from Intel. The electronics giant turned to Disney to showcase its next generation drones.

Video captured from drones has become the must-have money shot when selling high-end real estate. Equipped with a camera, the unmanned aerial vehicles soar over homes, giving prospective buyers a bird's-eye view of the property along with sweeping vistas and skylines. Recently, however, intrepid agents and drone pilots have turned them loose inside estates to capture the majesty of the great indoors: exhilarating descents past chandeliers that continue through cavernous living rooms before launching viewers straight out windows to behold cityscapes. "If you've got the space, using drones inside can be magical," Realtor Bret Parsons said. "One client said that they never realized how beautiful their home was until it was shot by a drone. And now they joke that they don't want to sell it."