[Congressional Record Volume 169, Number 123 (Tuesday, July 18, 2023)]
[Senate]
[Pages S2971-S2972]
From the Congressional Record Online through the Government Publishing Office [www.gpo.gov]
Artificial Intelligence
Mrs. BLACKBURN. Mr. President, AI is really grabbing some headlines,
and over the past few months, Tennesseans have been watching the Senate
as we go about having hearings and deciding how to handle this emerging
technology. This technology has broken boundaries in ways that most
people never imagined would be possible, and they see great potential
in it.
Healthcare companies in Tennessee are already using AI to enhance the
quality of care that doctors can offer their patients. If you have been
in for a doctor's checkup in the past few years, you have probably
noticed that your doctors and nurses are using more digital tools to
help inform their diagnoses and choose which drugs they are going to
prescribe. Most of those tools are powered by AI, as are many of the
apps and gadgets that healthcare providers recommend to their patients
to monitor their health and keep their treatments on track. We have
seen the use of these remote monitoring tools just explode, especially
in rural areas during the pandemic.
We are also seeing AI make a substantial impact on manufacturing. The
largest auto manufacturers in Tennessee are already using this
technology to make our cars safer and more efficient. Many of them are
experimenting with automated driving systems. They are also using it to
make operating a vehicle a little more practical. One that I found
amusing is that Ford used AI to develop a program that would
automatically hook a trailer to a hitch. Now, I know some of our
Members who have farms and ranches probably think you should have that
skill, but I think, for many of us, that would be a practical
application.
So whether we are talking about making things safer, finding which
pharmaceuticals are going to better treat diseases, or making things
more convenient, AI has the potential to completely transform entire
industries. Indeed, innovators are working on those components as we
speak.
It also comes with significant drawbacks, and nowhere are those
problems more pronounced than in Tennessee's music industry. All things
considered, technology has been great for the business side of the
music industry. The increased use of social media by up-and-coming
musicians has allowed record labels to sign artists who have millions
of fans. Platforms like Spotify automatically generate playlists that
promote new music, and all of that music is available instantly via an
app on your cell phone. But the thousands of musicians, songwriters,
and artists who call Tennessee home are really quite worried about the
lack of control they have over how AI platforms exploit their creations
and their intellectual property.
This month, the Senate Judiciary Committee held a hearing to examine
the downsides of making technology an essential element of the creative
process. The most obvious problem that we have uncovered in our
discussions with artists and songwriters comes down to a basic question
of copyright infringement.
We know that to create artistic works, AI systems must train on data
scraped from existing artistic works. If you look at how this
technology has advanced, it becomes clear that generative AI is often
trained on popular music that was used without the copyright holder's
permission. One AI model that we reviewed was trained on a staggering
280,000 hours of music. There is not a reasonable person who would ever
suggest that the writers and artists were compensated fairly for that
280,000 hours of use of their products that was used to train AI to
mimic and clone their voices and their words.
Now, the very existence of this tech and what it produces are enough
to draw challenges to spurious claims that data scraping counts as what
is called fair use, but we have even more extreme examples of what AI
can do after training on existing popular music, or music that has been
scraped, in order to train the platform.
Most recently, an anonymous content creator uploaded an AI-generated
song called ``Heart on My Sleeve.'' It was a perfect replication of the
vocals of Drake and The Weeknd. It was so perfect that listeners on
multiple platforms were duped into thinking it was real. The song
gained over 9 million views on social media and hundreds of thousands
of plays on Spotify and YouTube before the platforms took it down. But
because the internet is forever, so is this fake song, and there is
nothing those artists can do about it.
Not only does this raise questions about what ``fair use'' means in
the age of AI, but it also raises serious concerns about our current
State-by-State approach to protecting an individual right of publicity.
Common sense tells us that this isn't right, even if the law is still
catching up with the pace of technology. We have discussed having a
Federal--a Federal--individual right of privacy law.
AI has also affected how fans interact with real music. At least on
streaming platforms, playlists have outpaced platforms as the preferred
listening format. Companies have employed AI-driven playlists to expose
listeners to new artists, with decidedly mixed results.
Martina McBride, who knows a few things about country music, has gone
public about her experience in trying to build a playlist using
Spotify's ``recommender'' system. In a playlist she prompted with the
words ``country music,'' she had to refresh her recommendations 13
times before it delivered a song by a female artist. Country music is
an expansive genre. Yet, somehow, that algorithm spit out 120 songs by
men before managing to include a song by a woman.
We could chalk up this apparent bias to a glitch in evolving tech
were it not for the fact that fans rely almost exclusively on
technology to listen to and discover new music. People expect more
innovation, not less, because that is exactly what Big Tech is selling
them. When they spin one of those customized playlists, they expect to
hear a fair sampling of whatever it is they asked for, but that is
clearly not what these platforms are delivering. This matters to
established artists like Martina, but it matters even more to up-and-
coming artists who are forced to rely on these algorithms to gain
exposure to the music-listening public.
Over the past few years, Congress has learned a very important lesson
about the perils of allowing technology to outpace the law.
Over the course of several hearings, we produced evidence of the
unraveling
[[Page S2972]]
of safety and security standards on social media platforms. Senator
Blumenthal and I worked together on this. We heard testimony from
parents, children, and industry experts. We exposed how platforms
prioritize growth over maintaining even the most basic checks on
predatory and dangerous content.
This Congress, we have the opportunity to put guardrails on these
companies through the passage of the Kids Online Safety Act. It is a
strong, bipartisan bill that will make a real difference for children
and their families.
I would like to remind my colleagues that it took us far too long to
engage on this issue. Indeed, when the Acting President pro tempore and
I were in the House, we were leading the way in getting our colleagues
to address these issues.
I think we are at the point that we can say now that this has to be
done. We can't afford to wait and let AI systems be put in place
without having protections for the American people. So, in recent
months, we have held several hearings on this issue. Most recently, we
examined the impact on intellectual property rights. In June, we
fielded several firsthand accounts of how this technology could be
weaponized against basic human rights. Senator Ossoff and I led that
hearing.
We have yet to truly dig into the dangers of allowing China to
outpace us through the use of AI, and I think it is time that we catch
up.
When it comes to tech, we have done our best work when we have
listened to the people and have worked with people who are closest to
the issue. The Kids Online Safety Act came to be because we listened
when parents spoke out about how impossible it was to protect their
children in the virtual space.
Years ago, when the House Energy and Commerce Committee was crafting
what would become the 21st Century Cures Act, I worked with Tennessee's
healthcare innovators to include provisions that made the law more
responsive to the rapid growth of the healthcare technology industry.
We had similar conversations in the runup to the passage of the Music
Modernization Act.
So I would encourage my colleagues to start engaging with us on this
issue before it gets away from us. Listen to what creators and
engineers have to say about AI's effects on their industries.
There is no turning back the clock on AI just as we learned that
there was no turning back the clock on social media and that there was
no turning back the clock on the World Wide Web. AI is here to stay--it
is going to drive innovation--but that doesn't mean we have to accept
innovation at any cost.
I yield the floor.
The ACTING PRESIDENT pro tempore. The Senator from Michigan.
Ms. STABENOW. Mr. President, first, I just want to indicate also that
there is a lot to think about here when we are looking at artificial
intelligence. There are a lot of very serious questions we have to
figure out.
I am really pleased that we are seeing this bipartisan initiative in
the Senate, with Senator Schumer, our majority leader, coming forward,
but also with other colleagues who are really helping us go through
briefings and are really looking at all of the questions, because it is
really serious in terms of the questions. It is something we have to
deal with as this is not going away. Certainly, there is a lot we have
to figure out in all of this.