London metropolitan police
A report from the London Assembly police and crime committee suggests that the London Metropolitan Police force could add over 500 additional officers to their ranks and save up to £30 million with the implementation of AI technologies. The study urges an AI-adoption strategy similar to the NYPD, which is able to connect officers with live feeds from 8,000 city-wide CCTV cameras. Integrating this type of software has enabled the NYPD to cross reference information across various databases and surveillance systems. Additional use cases for AI-assisted law enforcement include identifying theft patterns and potential suspects, as well as supporting legal proceedings by analyzing complex digital paper trails, such as emails and text messages. Steve O’Connell, chairman of the London Assembly police and crime committee, believes that an investment in AI would empower the Metropolitan Police to solve crimes more efficiently, thereby increasing police performance. Artificial intelligence initiatives could also provide the department with better equipment for sifting through increasing amounts of digital evidence. O’Connell suggests the funds saved from investing in AI could then be used to employ hundreds of new officers.
Read more about the report at AI Business.
MIT drug research
MIT researchers are developing a fully automated molecule-design process with potential to develop faster, more consistent pharmaceutical products. Drug discovery depends on lead optimization—a process wherein chemists target a “lead” molecule. This molecule reacts with a specific biological system to produce the desired pharmaceutical effect. Each step of this process requires manual tweaking to match the lead molecule with the appropriate system. MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and Department of Electrical Engineering and Computer Science (EECS) have created an artificial intelligence model that selects optimal lead molecules. In successful use cases, these AI-selected molecules are expected to have the highest probability of matching effectively with the appropriate biological system. An effective match produces the desired pharmaceutical effect for drug researchers. The AI model is also trained to modify the drug’s molecular structure while ensuring the system remains chemically valid.
Read more about the project at MIT News.
AI art at auction
For the first time ever, an auction house plans to sell an AI-produced work of art. The piece was produced by an algorithm developed by the French art collective Obvious. Christie’s New York aims to auction off the work on canvas. The AI-created painting raises questions around whether an algorithm can be considered creative. The philosophical approach to the artistic processes is also challenged by the project’s success. Created via a Generative Adversarial Network (GAN), the artists fed their computer a dataset of 15,000 portraits painted between the 14th and 20th centuries. The algorithm then created new works based on the training set. The experiment was considered a success when one of the portraits passed a test designed to pinpoint whether a painting was made by a human or a machine. Titled Portrait of Edmond de Belamy, the work showcases a man clad in a dark coat and white collar. The piece is expected to bring an estimated $7,000 to $10,000. The art collective plans to dedicate proceeds from the sale to continue training its algorithm and expand its experiments to include AI-powered 3D modeling.
Read more about the auction at artnet News and Christie’s.
—
If you enjoyed our roundup about Artificial Intelligence and Machine Learning Industry News: London Metropolitan Police, MIT Drug Research, and AI Art at Auction, subscribe to our YouTube channel for video updates. 
