Microsoft today announced the launch of Anti-Corruption Technology and Solutions (ACTS), an initiative it says will leverage cloud computing, data visualization, AI, machine learning, and other technologies over the next decade to “enhance transparency” and “detect and deter” corruption. With ACTS, Microsoft says it hopes to “bend the curve” of corruption by helping governments innovate with expertise and other resources.
The United Nations (UN) reports that the cost of corruption is more than $3.6 trillion dollars a year, and global events of this year have created a world particularly vulnerable to corruption. As noted by the UN secretary-general Antonio Guterres, “Corruption … is even more damaging in times of crisis as the world is experiencing now with the Covid-19 pandemic. The response to the virus is creating new opportunities to exploit weak oversight and inadequate transparency, diverting funds away from people in their hour of greatest need.”
Microsoft says that over the last six months, it’s begun to make investments in support of ACTS, including a partnership with the Inter-American Development (IDB) Bank to advance anti-corruption, transparency, and integrity objectives in Latin America and the Caribbean. The company has also partnered with the IDB Transparency Fund to help bring greater transparency to the use of COVID-19 economic stimulus funds.
Separately, Microsoft researchers have worked with The World Bank to investigate the use of AI in anti-corruption efforts. The two conducted experiments with datasets from international organizations, national procurement data, beneficial ownership, and other corporate databases to detect patterns that hint at the possibility of corrupt behavior, exposing things like links in bidding patterns during contract negotiations and ownership information from around the globe. Microsoft corporate vice president and general counsel Dev Stahlkopf asserts that AI can enable stakeholders to better map networks of relations, locations, use of shell companies, off-shore jurisdictions, and banking information to address potential risks before a contract is issued.
“At Microsoft, we believe corruption is an urgent global issue that can and must be solved,” Microsoft corporate vice president and general counsel Dev Stahlkopf wrote in a blog post. “We know, for instance, that data can illuminate hidden patterns and relationships to provide governments with better tools to ensure public moneys go to their intended purposes. Technology resources such as cloud computing, data visualization, artificial intelligence, and machine learning provide powerful tools for governments and corporations to aggregate and analyze their enormous and complex datasets in the cloud, ferreting out corruption from the shadows where it lives, and even preventing corruption before it happens.”
It’s unclear what concrete solutions might emerge from ACTS in the near term, and Microsoft, it’s worth noting, has itself been accused of contributing to corruption. In August 2018, the U.S. Justice Department and the Securities and Exchange Commission investigated the company for possible bribery in its pursuit of software sales in Hungary. Microsoft paid the agencies $25 million to settle the probe out of court.
That said, Microsoft isn’t the first to propose using technologies like AI in the fight against government corruption. In the U.K., Exiger and Transparency International (TI) are working to improve the latter’s capacity to analyze public records toward identifying risk for corruption. In Ukraine, the local chapter of TI has developed its own AI tool to reveal fraudulent bids in public procurement. The Brazilian Office of the Comptroller General has deployed a machine learning model to estimate the risk of corrupt behavior among its civil servants. And IBM has been working with the government of Kenya to algorithmically identify the primary drivers of bribes.
The launch of ACTS follows the debut of Microsoft’s election verification tool — ElectionGuard — and M365 for Campaigns.
Author: Kyle Wiggers
Source: Venturebeat
