HB 306-AI, DEEPFAKES, CYBERSECURITY, DATA XFERS  3:03:55 PM CHAIR SHAW announced that the first order of business would be HOUSE BILL NO. 306, "An Act relating to artificial intelligence; requiring disclosure of deepfakes in campaign communications; relating to cybersecurity; and relating to data privacy." 3:04:18 PM JOSH APPLEBEE, House Majority Staff, Alaska State Legislature, provided a sponsor statement for HB 306, [included in the committee packet] which read as follows [original punctuation provided]: HB 306 establishes policy for Artificial Intelligence (AI) use by state agencies to ensure transparent and fair practices to protect Alaskans as well as requires disclosure on political deepfakes to inform Alaskans when campaign information is altered with the intention to mislead. With the increasing availability of AI to streamline work processes, transparency, and guidelines for the use of AI by state agencies is paramount to ensure appropriate deployment and to protect Alaskans from potential harm. By requiring human oversight/responsibility for AI, a publicly posted inventory of AI in use by state agencies, and regular impact assessments that include a review and public report of benefits, liabilities, risks, accountability mechanisms in place, decision appeal processes, and effects on the liberty, finances, livelihood and privacy interest of individuals among other items HB 306 will ensure AI is working for the people of Alaska appropriately. HB 306 ensures the public will know when AI is employed by state agencies regarding the collection and processing of personal data and the generation of decisions. The bill provides important and necessary safeguards but also methods of recourse if the use of AI causes harm to an individual. Decisions pertaining to hiring practices, eligibility for services, license qualification, and more, are no small matters; they impact lives. AI use will also be regulated in relation to "synthetic media" or "political deepfakes" disseminated to the public, including for election and campaign use. The generation of audio, still-imagery, or video that is designed to create a fabricated or intentionally manipulated account of an individual's appearance, speech, or conduct is a fairly new phenomenon. When these digital alterations of a person, however, are employed to spread false information with malevolent intentions to deceive the public, it is time for lawmakers to take up the matter. With easily accessible AI available to develop deepfakes, the likelihood of their creation during the 2024 election cycle is inevitable. HB 306 requires a clear disclosure on a political deepfake that the communication has been manipulated by AI. AI is the new digital frontier, and as the rapidly emerging technology can offer efficiencies and solutions for the workplace and ease of use even for non-tech savvy individuals, we will see it play an increasing role in work activities and in political communications. The goal of HB 306 is to set appropriate parameters for state agency use of AI and to protect Alaskans. MR. APPLEBEE reported that artificial intelligence (AI) lobbying has increased 185 percent between 2022 and 2023 with over 450 organizations participating. He said that HB 306 was drafted to safeguard against bad actors and gain benefits of what artificial intelligence (AI) technology has to offer. Transparency and guidelines for AI usage are needed to ensure appropriate deployment to protect Alaskans from potential harm. He said HB 306 would create human oversight of AI and create a publicly posted inventory of AI use by state agencies and manage regular impact assessments including review and public report of benefits, liabilities, risks, and AI's affects on the privacy interests of individuals. MR. APPLEBEE explained that HB 306 would employ methods of recourse if the use of AI causes harm to an individual. He said HB 306 requires the disclosure of political "deepfakes" when campaign information is altered by AI technology. The Federal Communications Commission (FCC) recently made AI generated voices and robocalls illegal. With accessible AI tools, it has never been easier to create "deepfakes," and the likelihood of their creation during the upcoming election cycle is inevitable. MR. APPLEBEE reported that many major technology ("tech") companies signed a pact to voluntarily adopt reasonable precautions to prevent AI tools from being used to disrupt democratic elections around the world. U.S. Senator Lisa Murkowski introduced a pact to require disclaimers on political ads with images, audio, or video, that are substantially generated by AI. Similarly, he said, HB 306 would require a clear disclosure that demonstrates that the communication has been manipulated by AI. 3:08:45 PM SEAN MILLS, House Majority Staff, Alaska State Legislature, provided information about how AI is regulated in other states. He said 18 states enacted legislation associated with AI in 2023. He exemplified Connecticut, which requires an inventory of all AI systems that state agencies utilize. Some states enacted an AI advisory council. The Council of State Governments identified six focus areas for legislation, which include accountability, fighting discrimination, providing transparency for state operations, providing data privacy, targeting unsafe and ineffective AI systems, and providing for interagency collaboration. He said that HB 306 addresses each of those six points. He said there are many risks and benefits to AI. The proposed legislation provides a launchpad for valuable discussion. 3:12:04 PM CHAIR SHAW asked about data "xers". MR. MILLS answered that "xers" is shorthand for transfers. 3:12:32 PM REPRESENTATIVE CARRICK asked if the bill would cover state agency generated deepfakes or system generated deepfakes from private entities. MR. APPLEBEE responded that the bill would address two parts: one that handles deepfakes within the state, and another for those generated outside the state for campaign purposes. REPRESENTATIVE CARRICK asked about the cost of the inventory for state agency's use of AI for consequential decisions. MR. APPLEBEE said the fiscal note includes an initial cost of $5 million, and recurring costs of $2 million per year. 3:14:05 PM REPRESENTATIVE STORY noted the importance and timeliness of the bill. She asked for the definition of the term "consequential decision" used in the bill. MR. APPLEBEE referred to page 5, line 21, which states "consequential decision means a conclusion decision or judgement by a state agency that can have a legal or significant effect on an individual". REPRESENTATIVE STORY asked for a list of state agencies that may be involved with consequential decisions. MR. APPLEBEE stated that HB 306 would cover any agency that utilizes AI technology. The bill would require the Department of Administration (DOA) to catalog those uses. The fiscal note assigns 15 programmer positions that would cover each state agency. REPRESENTATIVE STORY asked if all agencies would be included. MR. APPLEBEE said if the agency decides to use AI technology to make decisions, it would be cataloged. REPRESENTATIVE STORY asked how cyber security is tracked and AI is used currently in state departments. MR. APPLEBEE responded that he was unsure because there has yet to be an inventory. He offered to reach out to DOA and follow up with more information. 3:16:43 PM REPRESENTATIVE CARRICK asked how frequent state generated deepfakes are and if they are prevalent elsewhere. MR. APPLEGATE made a distinction between the use of AI technology and the creation of deepfakes or synthetic media. He explained that it is unlikely that a state agency would be creating deepfakes. State agencies may be utilizing AI technology to make systems more efficient. 3:18:09 PM CHAIR SHAW announced that HB 306 was held over.