03/13/2024 01:00 PM House JUDICIARY

           HB 358-PROHIBIT AI-ALTERED REPRESENTATIONS                                                                       
                                                                                                                                
1:05:13 PM                                                                                                                    
                                                                                                                                
CHAIR VANCE announced  that the first order of  business would be                                                               
HOUSE  BILL  NO. 358,  "An  Act  relating  to use  of  artificial                                                               
intelligence to create or alter  a representation of the voice or                                                               
likeness of an individual."                                                                                                     
                                                                                                                                
1:05:31 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE  MIKE  CRONK,   Alaska  State  Legislature,  prime                                                               
sponsor,  presented HB  358.   He  shared  the sponsor  statement                                                               
[included  in  the  committee  packet],  which  read  as  follows                                                               
[original punctuation provided]:                                                                                                
                                                                                                                                
     In  this high-tech  world  Artificial Intelligence  has                                                                    
     become a lightning rod for  debate. HB   358 is written                                                                    
     as  simply as  possible to  address  the use  of AI  to                                                                    
     create  false  identities  and  cause  harm.  Literally                                                                    
     everything  we see  and  hear now  can  be called  into                                                                    
     question.  Is it  real,  is  it genuine,  or  is it  AI                                                                    
     created or  enhanced. Your voice, your  image will only                                                                    
     be yours and safe from  harm if safeguards are put into                                                                    
     place.                                                                                                                     
                                                                                                                                
1:06:20 PM                                                                                                                    
                                                                                                                                
DAVE STANCLIFF,  Staff, Representative  Mike Cronk,  Alaska State                                                               
Legislature, on behalf of Representative  Cronk, prime sponsor of                                                               
HB  358,  presented  the  sectional  analysis  [included  in  the                                                               
committee packet],  which read  as follows  [original punctuation                                                               
provided]:                                                                                                                      
                                                                                                                                
     Section 1. AS 45.50 is amended adding a new section.                                                                       
                                                                                                                                
     Article   6A.  Use   of   Artificial  Intelligence   to                                                                    
     Represent an Individual.                                                                                                   
                                                                                                                                
     Sec. 45.50.860 Unauthorized  representation of voice or                                                                    
     likeness.                                                                                                                  
                                                                                                                                
        • Restrictions in the use of AI to create or alter                                                                      
          a person's voice or likeness.                                                                                         
        • Sets the standard in the restriction of "Intent                                                                       
          to cause harm."                                                                                                       
        • Defines:                                                                                                              
          1. "Artificial Intelligence"                                                                                          
          2. Individual"                                                                                                        
          3. "Likeness representation"                                                                                          
                                                                                                                                
1:08:37 PM                                                                                                                    
                                                                                                                                
MR.   STANCLIFF  shared   several   examples   of  AI   generated                                                               
impersonations  of elected  officials.   One example  involved an                                                               
impersonation of President Biden and  impacted 25,000 people.  He                                                               
said the  proposed legislation would convey  that the legislature                                                               
is aware of this problem.                                                                                                       
                                                                                                                                
1:10:49 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE  GRAY informed  the  committee that  as a  podcast                                                               
host, he  would often  edit the order  of guests'  statements for                                                               
clarity.  He  sought to verify that the bill  would only cover AI                                                               
generated material.                                                                                                             
                                                                                                                                
MR. STANCLIFF confirmed.                                                                                                        
                                                                                                                                
REPRESENTATIVE  GRAY  opined   that  rearranging  people's  words                                                               
shouldn't be  allowed because  it would be  easy to  insert words                                                               
and  reverse the  meaning of  sentences.   He suggested  that the                                                               
amendment should be broader in scope.                                                                                           
                                                                                                                                
MR. STANCLIFF  said that  would make  the bill  more complicated,                                                               
and the  goal was  to keep  it narrow and  focused.   He stressed                                                               
that the use of this  technology in child pornography is rampant.                                                               
He indicated  that the intent  to cause harm would  be determined                                                               
by the courts.                                                                                                                  
                                                                                                                                
1:14:44 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE CARPENTER  questioned the constitutional  right of                                                               
free speech when applied to the use  of AI and how to ensure that                                                               
AI is not unduly influencing.                                                                                                   
                                                                                                                                
MR. STANCLIFF suspected that if  the intent of using AI generated                                                               
images was to cause harm, that  person could be held liable under                                                               
the proposed legislation.                                                                                                       
                                                                                                                                
REPRESENTATIVE CARPENTER  sought to  confirm that the  bill seeks                                                               
to modify AS 45.50.                                                                                                             
                                                                                                                                
MR. STANCLIFF answered yes.                                                                                                     
                                                                                                                                
REPRESENTATIVE CARPENTER  asked whether there was  any discussion                                                               
with Legislative Legal Services about the definition of harm.                                                                   
                                                                                                                                
MR. STANCLIFF  stated that the definition  was kept intentionally                                                               
broad.                                                                                                                          
                                                                                                                                
1:18:09 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE SUMNER  pointed out  that the  proposed definition                                                               
of AI contemplates a machine-based  system with varying levels of                                                               
autonomy.   He asked  whether the bill  sponsor had  considered a                                                               
scenario in which  the AI system has complete  autonomy and asked                                                               
who would be liable.                                                                                                            
                                                                                                                                
MR. STANCLIFF pointed  out that AI is widely used  for many great                                                               
purposes.   He  reiterated  that liability  would  depend on  the                                                               
intent to cause harm.                                                                                                           
                                                                                                                                
1:20:44 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE GROH  asked whether this issue  had been addressed                                                               
in other states.                                                                                                                
                                                                                                                                
MR. STANCLIFF  shared his  belief that at  least 35  other states                                                               
are delving into  this in one way or another.   In addition, deep                                                               
fakes are being addressed by  national legislation and technology                                                               
companies are being asked to be more responsible.                                                                               
                                                                                                                                
REPRESENTATIVE  GROH  questioned  the  origins  of  the  proposed                                                               
damages of  up to $2,000  for the  first violation not  to exceed                                                               
$2,500 for second or subsequent violations.                                                                                     
                                                                                                                                
MR. STANCLIFF said  the bill sponsor was open  to the committee's                                                               
suggestions.                                                                                                                    
                                                                                                                                
REPRESENTATIVE GROH addressed the concept  of intent.  He pointed                                                               
out that in  most scenarios, both parties  typically believe they                                                               
are doing  the right thing.   He asked how that  kind of thinking                                                               
played into the intent to cause harm standard.                                                                                  
                                                                                                                                
MR. STANCLIFF  commented on the  ability to convince a  judge and                                                               
jury that someone was harmed [by the use of AI].                                                                                
                                                                                                                                
REPRESENTATIVE  GROH asked  whether [the  proposed damages]  were                                                               
enough.                                                                                                                         
                                                                                                                                
MR. STANCLIFF reiterated his hope  that the committee would delve                                                               
deeper  into  these issues.    He  characterized  the bill  as  a                                                               
starting point.                                                                                                                 
                                                                                                                                
1:26:30 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE CARPENTER clarified that  he was supportive of the                                                               
bill's intent.  He asked whether  the definition of AI matters if                                                               
its purpose  is to cause harm.   In addition, he  asked who would                                                               
be responsible when the AI's response is counter to its intent.                                                                 
                                                                                                                                
MR. STANCLIFF  responded, "If  your finger  hit the  keyboard, it                                                               
could come back to  bite you."  He reiterated that  if the use of                                                               
AI  is  intended to  cause  harm,  bully  someone, or  destroy  a                                                               
business, the bill would open up  the door for more efficient and                                                               
justifiable civil action.                                                                                                       
                                                                                                                                
1:30:13 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE SUMNER suggested  that an AI program  of a certain                                                               
level  of  autonomy  could  be  [defined as]  a  person  for  the                                                               
purposes of liability under the proposed legislation.                                                                           
                                                                                                                                
CHAIR  VANCE   clarified  that  the   bill  would   create  civil                                                               
liability, not criminal.                                                                                                        
                                                                                                                                
REPRESENTATIVE  SUMNER  pointed  out   that  in  terms  of  civil                                                               
liability,  incarcerating   a  computer  program  would   not  be                                                               
entirely possible.                                                                                                              
                                                                                                                                
1:32:04 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE  GRAY gave  an  example  involving two  candidates                                                               
similar in race, age, gender, and  height.  One of the candidates                                                               
presents  himself/herself as  5  inches taller  with  the use  of                                                               
Photoshop  and  wins  the  race.   Considering  that  the  taller                                                               
candidate tends  to win in  presidential races, he  asked whether                                                               
the use of Photoshop would be  considered AI and whether harm was                                                               
done.                                                                                                                           
                                                                                                                                
MR. STANCLIFF answered no.   He reiterated that the bill pertains                                                               
to AI that's  used to enhance or change an  image with the intent                                                               
to  cause damage.   In  contrast,  if AI  was used  to make  your                                                               
opponent appear shorter  or to alter his/her speech  in a strange                                                               
way, then the bill might apply.   He maintained that because it's                                                               
a civil matter,  the courts would need to make  a judgement based                                                               
on intent.                                                                                                                      
                                                                                                                                
1:34:37 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE  GRAY  considered  a  scenario in  which  a  radio                                                               
station unintentionally uses  a clip of President  Biden that was                                                               
AI generated  and asked whether  the radio station would  be held                                                               
liable.                                                                                                                         
                                                                                                                                
MR.  STANCLIFF  suggested  that  the  committee  should  consider                                                               
indemnifying people who  can present an honest  answer for airing                                                               
AI generated  content.  He pointed  out that AI has  already been                                                               
used  in politics.    He  emphasized that  the  bill  would be  a                                                               
starting point for identity protection.                                                                                         
                                                                                                                                
REPRESENTATIVE  GRAY opined  that $2,000  is not  a deterrent  to                                                               
media companies.                                                                                                                
                                                                                                                                
REPRESENTATIVE CARPENTER stated that  intent matters.  He likened                                                               
AI generation  to an uncontrollable forest  fire, indicating that                                                               
regardless of intent,  the person who "pushed  the button" should                                                               
be held responsible.                                                                                                            
                                                                                                                                
1:38:57 PM                                                                                                                    
                                                                                                                                
REPRESENTATIVE   C.  JOHNSON   pointed  out   that  the   Federal                                                               
Communications  Commission (FCC)  has  very stringent  guidelines                                                               
that does not include the use of AI.                                                                                            
                                                                                                                                
REPRESENTATIVE  GRAY  stated,  "One  person's  spoof  is  another                                                               
person's truth,"  and shared his  belief that  [the use of  AI on                                                               
the radio] is something to be concerned about.                                                                                  
                                                                                                                                
REPRESENTATIVE  GROH suggested  that political  satire or  parody                                                               
should not be sanctioned due  to First Amendment protections.  He                                                               
emphasized  the difficulty  of this  task given  the novelty  and                                                               
importance of the topic.                                                                                                        
                                                                                                                                
1:43:37 PM                                                                                                                    
                                                                                                                                
CHAIR VANCE announced that HB 358 would be held over.