Sen. Warner: AI firms should put security at the center of their work

The top Democrat on the Senate Intelligence Committee wants answers to questions ranging from supply chain security to privacy.

The top Democrat on the Senate Intelligence Committee is pushing leading artificial intelligence companies to place greater emphasis on addressing security concerns posed by the technology as it rapidly advances.

In a set of letters sent Wednesday to the CEOs of top AI firms, Sen. Mark Warner, D-Va., urged the companies to put “security at the forefront of your work.” By embracing security-by-design principles, Warner noted, companies have the chance to address the harmful consequences of their technologies today, rather than further down the road — as has been the case too often in the recent history of technological development.

And as AI becomes integrated into a wide swathe of applications, “it is imperative that we address threats to not only digital security but also threats to physical security and political security,” Warner wrote.

The letters probe the companies on several security matters, including how they are protecting systems from attacks injecting bad data into models, their process for monitoring and auditing to detect data breaches and unauthorized use and how they have handled past security incidents. The recipients include OpenAI, Scale AI, Meta, Google, Apple, Stability AI, Midjourney, Anthropic, Percipient.ai and Microsoft. The letters request answers to a set of detailed questions about the firms’ security practices no later than May 26.

Several recipients did not immediately return requests for comment. A spokesperson for Anthropic said the company thanked the Senator for his interest and intended to respond in “due course.”

To emphasize his point, Warner’s staff used AI to write the majority of the press release. The Virginia Democrat noted on Twitter the AI did “pretty darn well.”

While cybersecurity risks posed by AI, such as the rapid creation of new malicious code, have so far been largely hypothetical, the security of the companies producing the technology has been called into question on several occasions. For instance, in March Meta’s language model, known as LLaMA, was leaked early online.

Warner also asks the companies to provide more information on concerns related to trustworthiness, researcher transparency and potential algorithmic bias. The letter follows a joint announcement Tuesday from four major U.S. enforcement agencies committing to using existing U.S. regulations to crack down on AI products perpetuating discrimination and fraud.

While firms like OpenAI — the industry leader — have made commitments to address the security and societal concerns posed by its products, Warner, an influential voice on tech policy in the Senate, said this did not obviate the need for more stringent rules. “Beyond industry commitments, however, it is also clear that some level of regulation is necessary in this field,” Warner wrote in the letters.

The risks posed by AI have caught the attention of other top lawmakers. Earlier this month, Senate Majority Leader Chuck Schumer, D-N.Y., announced he was working on a regulatory framework for AI in. European Union lawmakers recently called for a global summit on AI, citing the risks of use of the technology by non-democratic countries, and are considering a wide-ranging proposal to regulate use of the technology, the EU AI Act. Regulators in Italy temporarily banned OpenAI’s ChatGPT over privacy concerns.

100-plus companies call for ‘global surge’ in AI-powered cyber defense

OpenAI, Anthropic, Google, Microsoft, and others say there’s a narrow “defenders’ window” to strengthen security before AI-powered attacks become more sophisticated.

OpenAI: Agent behavior that led to Hugging Face intrusion formed in May

The push to designate AI as the next critical infrastructure sector

Government

• FBI officials say AI is bolstering adversaries, emphasizing need to focus on cyber basics, patching

• Feds accuse China of ‘systematic’ distillation of U.S. AI models

• CIA’s Michael Ellis says cyber intelligence is changing how the agency operates

• The G7 tells industry to hurry up and prep for post-quantum encryption

Technology

• European parliament members call for slowdown of Serbia’s EU entry over spyware use

• FCC proposes public scorecard to rate telecoms on anti-robocall efforts

• Pegasus, NoviSpy variant spyware found on devices of Serbian activists

• Wyden seeks upgraded NSA security guidance on commercial VPN use

Threats

• Chinese espionage groups swarm to exploit triple-link chain of zero-days

• Microsoft discloses two actively exploited zero-days among 974 vulnerabilities

• Russian national extradited to US for alleged involvement in bank-account takeover scheme

• Attackers exploit zero-days in consistently besieged SonicWall product

Policy

• Whistleblower says USPS deploying new, ‘untested’ IT systems governing mail-in ballots

• ‘Watershed 250’ test program in Texas looks to private sector for water cybersecurity help

• Former sexual abuse victims say Grok used their images, videos to train deepfake capabilities

• Cyber threats nudge Trump to sign executive order on foreign equipment in U.S. energy infrastructure