Gov. Hochul bans DeepSeek AI from state government computer systems

Gov. Kathy Hochul delivers her State of the State speech at the Empire State Plaza Convention Center in Albany on Jan. 14. Credit: Newsday via Getty Images/Newsday LLC
ALBANY — Gov. Kathy Hochul on Monday banned the DeepSeek software from state computer systems over concerns the Chinese artificial intelligence product could steal data and technology secrets in a cyberattack.
Hochul’s move comes as academic researchers warn of the potential capabilities of DeepSeek to mount cyberattacks in the United States. Government officials claim the company that produced DeepSeek is linked directly to China.
A similar ban of DeepSeek is being pushed in Congress.
"We’re working aggressively to protect New Yorkers from foreign and domestic threats," Hochul said in a statement Monday. "New York will continue fighting to combat cyber threats [and] ensure the privacy and safety of our data, and safeguard against state-sponsored censorship."
WHAT NEWSDAY FOUND
- Gov. Kathy Hochul has banned the DeepSeek software from state computer systems over concerns the Chinese artificial intelligence product could steal data and technology secrets in a cyberattack.
- The move comes as academic researchers warn of the potential capabilities of DeepSeek to mount cyberattacks in the United States.
- A similar ban of DeepSeek is being pushed in Congress.
DeepSeek is AI that is capable of reasoning through problems like a human. It can be used in many ways, including computer coding. DeepSeek has attracted attention from U.S. companies seeking to capitalize on it, which many experts in the field consider better than other AI models and, according to the company that promotes it, far cheaper.
But concerns continue as to whether DeepSeek could extract data that could jeopardize personal, corporate and national security. New York’s computer systems hold extensive data, including New Yorkers’ personal, tax and health data; information from corporations that is proprietary and shielded from competitors, and public safety resources and strategies.
There is much to be concerned about with about DeepSeek, but also reason for interest in the potential promise of this new step in AI, said Rob Glew, a professor at McGill University in Montreal who has studied DeepSeek.
He said governments are concerned a public worker might use DeepSeek to help develop public programs, such as social services.
"These models aren’t static, they are data driven," Glew said in an interview with Newsday. "The data you put in is being used to advance, to provide more data for it to learn ... you typed in some details of an important welfare program ... but we don’t know where it’s going."
He said DeepSeek appears capable of taking some data from one user and using data from other users to secure information that was intended be protected within state computers.
Much of the concern stems from DeepSeek’s privacy policies, said Dominic Sellitto, assistant faculty director of the University at Buffalo Center for AI Business Innovation.
He said many AI producers have a policy that states: "Any data you give us, we may sell and share with third parties." Other AI producers address that concern by having different security levels for government and corporate versions of their products, he said. DeepSeek doesn’t appear to have a more secure government-level product at this time, he said. The result is bans by governments.
"I think it’s relatively in line with the what we are seeing around the world," Sellitto said.
Another concern is in spreading disinformation, such as how it describes issues that conflict with the goals of the Chinese government, or to provide an overly positive view of the government to the world.
"The classic thing is if you ask DeepSeek what happened at Tiananmen Square. It says, 'I don’t know about that,'" Glew said of the 1989 uprising against China's communist government. "There is evidence of that public disinformation ... which seems to be inspired by the manifesto of the Community Party."
But DeepSeek is also prompting other AI companies to find less expensive ways to provide the benefits of using AI to solve problems and develop ideas, Glew said.
The Washington contact for DeepSeek wasn’t taking questions from reporters.
Last week, a bipartisan bill was introduced into the House dubbed, "No DeepSeek on Government Devices Act." The sponsors said the ban is needed because of China’s ability to use the app for surveillance and misinformation.
"The Chinese Communist Party has made it abundantly clear that it will exploit any tool at its disposal to undermine our national security, spew harmful disinformation, and collect data on Americans," Rep. Josh Gottheimer (D-N.J.) said in a statement to The Associated Press.
Co-sponsor Rep. Darin LaHood (R-Ill.) said the federal ban would close "backdoor operations the company seeks to exploit for access."
DeepSeek has already been banned or partly banned for use by government in Texas, Italy, Australia, Taiwan and South Korea, according to the AP.
Suffolk inmates allege jail abuse ... Arrest in American Armor theft ... Babylon cooking lessons ... What's up on LI
Suffolk inmates allege jail abuse ... Arrest in American Armor theft ... Babylon cooking lessons ... What's up on LI