In short Governments around the world should pass intellectual property laws that grant rights to AI systems, argued two academics from the University of New South Wales in Australia.
Alexandra George and Toby Walsh, professors of law and artificial intelligence respectively, believe that failure to recognize machines as inventors could have lasting impacts on economies and societies.
“If courts and governments rule that inventions created by AI cannot be patented, the implications could be enormous,” they wrote in a commentary post. published in nature. “Funders and companies would have less incentive to pursue useful research using AI inventors when the return on investment might be limited. Society might miss out on developing worthwhile and life-saving inventions.”
Laws today pretty much only recognize humans as inventors with intellectual property rights protecting them from patent infringement. Attempts to overturn human-centered laws have failed. Stephen Thaler, a developer who insists that AI invented his company’s products, has unsuccessfully sued trademark offices in several countries, including the US and UK.
George and Walsh side with Thaler. “Creating bespoke law and an international treaty won’t be easy, but not creating them will be worse. AI is changing the way science is done and inventions are made. We need property law intellectual property to ensure that it serves the public well,” they wrote.
Dutch police generate deepfake of teen who died in criminal case
A video clip featuring the face of a 13-year-old boy, who was shot outside a metro station in the Netherlands, swapped to a body using AI technology, has been released by police .
Sedar Soares died in 2003. Officers failed to solve the case and, with permission from Soares’ family, they generated a deepfake of his image on a child playing football in a pitch, presumably to help jog anyone’s memory. Cops have reportedly received dozens of potential leads since, according to the Guardian.
This is the first time that AI-generated images have been used in an attempt to solve a criminal case, it seems. “We have not yet checked whether these tracks are usable,” said Rotterdam police spokeswoman Lillian van Duijvenbode.
You can watch the video here.
AI task force advises Congress to fund national IT infrastructure
The US National Artificial Intelligence Research Resource (NAIRR) has urged Congress to launch a “shared research cyberinfrastructure” to better provide academics with hardware and data resources to develop machine learning technology.
The playing field for AI research is uneven. State-of-the-art models often contain billions of parameters; developers need access to many computer chips to train them. This is why research in private companies seems to dominate, while academics lag behind.
“We need to ensure that everyone across the country has the capacity to pursue cutting-edge AI research,” NAIRR wrote in a report. “This growing resource divide has the potential to negatively skew our AI research ecosystem and, in doing so, threaten our nation’s ability to cultivate an AI research community and workforce that reflects the America’s rich diversity – and harnessing AI in a way that serves all Americans.”
If advances in AI are driven by private companies, it could mean that other kinds of research areas are left behind and underdeveloped. “Develop and diversify approaches and applications of AI and open up opportunities for progress in all fields and scientific disciplines, including in critical areas such as auditing, testing and evaluation of AI, trusted AI, bias mitigation, and AI safety,” the task force said. argued.
You can read the full report here [PDF].
Meta offers musculoskeletal research technology
Meta AI researchers have released Myosuite, a set of musculoskeletal models and tasks to simulate limb biomechanical movement for a range of applications.
“The more intelligent an organism, the more complex the motor behavior it can exhibit,” they explain. said in a blog post. “So an important question to consider then is – what enables such complex decision-making and the motor control to execute those decisions? To explore this question, we developed MyoSuite.”
Myosuite was built in collaboration with researchers at the University of Twente in the Netherlands, and aims to arm developers who study prosthetics and could help rehabilitate patients. However, there is another potential useful application for Meta: creating more realistic avatars that can move around the metaverse more naturally.
The models so far only simulate arm and hand movements. Tasks include using machine learning to simulate handling a dice or spinning two balls. The application of Myosuite in the metaverse of Meta is a bit ironic given that there is no touching allowed there with restrictions on hands to deter harassment. ®