Thu Apr-11-2024: Navigating the Copyright Minefield,AI Safety Research,AI Use Cases

[00:00:00] It's Thursday, April 11, 2024 and I'm Dave Sobel, three things to note today.

[00:00:07] Navigating the copyright minefield, Genitive AI's challenge to content creation, legislation

[00:00:14] targeting AI copyright compliance and introduces fines for lack of creator disclosures, and

[00:00:20] Kaseya Empower sales teams with Genitive AI Assistant slashing documentation time.

[00:00:26] This is The Business of Tech.

[00:00:30] In today's digital work environment, the browser is the most widely used operating system, yet

[00:00:35] remains a significant security blind spot.

[00:00:38] Adacama has introduced the browser security platform purpose-built for MSPs.

[00:00:44] Built in collaboration with over 40 MSPs, this solution focuses on managing and protecting

[00:00:49] today's window of work, the browser.

[00:00:52] Adacama's browser extension empowers MSPs to set data policies, gain critical insights

[00:00:58] and optimize the user experience.

[00:01:00] Adacama recently announced the limited time 50 for 50 program, discounting the per user

[00:01:05] price by 50% when you deploy the plugin to 50% or more of your client endpoints.

[00:01:12] Visit adacama.com slash msp radio to learn more and see the platform in action.

[00:01:18] That's A-T-A-K-A-M-A dot com slash MSP radio to learn more.

[00:01:24] When I take time off, I get the luxury of reviewing all the stories to bring you the most interesting

[00:01:30] ones so we don't miss anything.

[00:01:32] There are some good ones, so let's start with the AI News.

[00:01:37] Despite the hype around AI safety, a new study reveals that only 2% of AI research

[00:01:42] is dedicated to the topic.

[00:01:44] American institutions and companies contribute the most to AI safety research, but it

[00:01:48] still pales compared to the overall volume of AI studies.

[00:01:53] While the number of AI safety research papers has increased globally, there's debate about

[00:01:57] the definition of AI safety and its focus on existential risks.

[00:02:02] The study suggests that more funding and resources should be allocated to AI research to keep

[00:02:08] up with industry advancement.

[00:02:11] Anthropics Claw 3 Opus has surpassed OpenAI's GPT-4 on Chatbot Arena, marking a significant

[00:02:18] moment in the history of AI language models.

[00:02:21] This achievement is notable as GPT-4 has consistently held the top position on the leaderboard.

[00:02:28] The victory of Claw 3 Opus highlights the importance of diversity among top vendors in the AI space.

[00:02:33] The Chatbot Arena, run by a large model systems organization, plays a crucial role in measuring

[00:02:38] the performance of AI Chatbots.

[00:02:41] OpenAI is expected to release a new successor to GPT-4 Turbo later this year.

[00:02:47] I also want to highlight coverage in Nextgov, which looks at copyright concerns around AI.

[00:02:53] Generative AI tools have the potential to create content that may infringe on copyright protections.

[00:02:59] The challenge lies in how individuals and companies can be liable for copyright violations when

[00:03:04] using generative AI tools.

[00:03:06] Selective prompting strategies can result in unintentional copyright infringement,

[00:03:11] and there's a need to establish guardrails to prevent such violations.

[00:03:15] Methods for detecting copyright infringement, establishing content providence, and ensuring

[00:03:20] model trading reduces similarity to copyrighted material has been suggested.

[00:03:25] The article notes that policymakers and regulation may play a role in providing best practices

[00:03:30] for copyright safety, such as filtering or restricting model outputs and regulating

[00:03:35] dataset construction and model training.

[00:03:38] Why do we care?

[00:03:40] The value providers in the future will be in their expertise in guiding customers here.

[00:03:45] That will include understanding which models and products perform well for which scenarios,

[00:03:50] and the risks and concerns the product use will result in.

[00:03:54] That's why we care.

[00:03:56] In order to deliver those services, we'll need research and data so we should care

[00:04:00] about how much funding is provided there.

[00:04:02] And we'll get to laws in a moment.

[00:04:07] So let's review some legislation and start with things that have passed.

[00:04:11] Tennessee has passed a First In the Nation bill to ensure like this Voice and Image Security

[00:04:16] Act to protect musicians from artificial intelligence by penalizing the unauthorized copying of a

[00:04:22] performer's voice.

[00:04:24] The use of AI technology in mimicking public figures led to discussion around regulations,

[00:04:29] and this new law aims to safeguard the music industry in Tennessee, which generates

[00:04:33] billions of dollars for the state and supports thousands of jobs and venues.

[00:04:38] In policy but not law, the White House has introduced extensive AI guidelines for the

[00:04:43] federal government requiring agencies to stop using AI systems that infringe on Americans'

[00:04:49] rights or safety.

[00:04:52] Policy also emphasizes transparency and internal oversight of AI use, including

[00:04:56] appointing chief AI officers and establishing AI governance boards.

[00:05:01] The guidelines aim to ensure responsible AI adoption while harnessing its potential to

[00:05:06] improve public services and address societal challenges.

[00:05:10] The headliner in Bills That Have Been Introduced is a bipartisan U.S. federal data protection

[00:05:16] law called the American Privacy Rights Act, introduced by Congresswoman Cappy McMorris-Rodgers

[00:05:22] and Senator Maria Cantwell.

[00:05:24] The law aims to give U.S. citizens greater control over their personal data, strengthen

[00:05:29] cybersecurity standards, and prohibit discrimination based on personal information.

[00:05:34] It also seeks to eliminate the patchwork of state privacy laws and establish a national

[00:05:39] data privacy and security standard.

[00:05:42] A bill introduced in the U.S. House of Representatives would require those training AI models to

[00:05:48] disclose any copyrighted works used and it would apply retroactively.

[00:05:53] The bill aims to increase transparency and protect creators' rights in the age

[00:05:57] of AI.

[00:05:58] The bill also aims to ensure to comply with the disclosure requirement could result in

[00:06:01] a penalty of at least $5,000.

[00:06:04] California has introduced a Right to Disconnect Bill requiring employers to specify work

[00:06:09] hours and prohibit employees from responding to work-related communications all off

[00:06:14] the clock.

[00:06:15] Assemblyman Matt Hanney introduced a proposition to update protections for the modern year.

[00:06:21] In political moves, Republicans who voted against federal bills for broadband funding

[00:06:26] are now taking credit for the local broadband projects funded by those bills.

[00:06:31] Despite opposing the bills, politicians like Senator Tommy Tuberville, Governor Ron DeSantis,

[00:06:37] Governor Greg Giaforte, Senator John Corden, and Senator Shelley Moore Capito are promoting

[00:06:43] and celebrating the benefits of the broadband subsidies.

[00:06:47] This is despite the elimination of the Affordable Connectivity Program which

[00:06:51] announced reduced support amounts for May as the program winds down.

[00:06:55] Finally, law is taking effect.

[00:06:57] All but the smallest ISPs are now required to publish broadband nutrition labels on their

[00:07:03] plans, providing consumers with information on costs, fees, speeds, and more.

[00:07:08] The labels aim to make it easier for consumers to compare plans and avoid hidden fees.

[00:07:13] Major broadband providers have fought against this requirement while some critics argue

[00:07:17] it doesn't address the issue of regional broadband monopolies.

[00:07:21] Verizon, Google Fiber, and T-Mobile have already released labels ahead of the deadline.

[00:07:27] Why do we care?

[00:07:29] Tennessee moves to protect its industry.

[00:07:31] As a producer of music, they distinctly care about this risk.

[00:07:35] The headline here is the possibility of a U.S. federal level privacy law.

[00:07:40] This is bipartisan, so that's a positive.

[00:07:43] Congress has limited time to legislate, so we'll see what happens here.

[00:07:47] I'm going to observe that companies should have a policy about requirements for work hours

[00:07:52] and expectations for after-hours communications, particularly in the IT space.

[00:07:58] Does it need a law?

[00:08:00] Regardless, it's a wise investment for savvy owners looking for shire employee retention.

[00:08:08] Let's also cover some AI use cases.

[00:08:12] Texas is implementing an automated scoring engine powered by artificial intelligence

[00:08:17] to grade open-ended questions on state exams, aiming to replace a majority of human graders.

[00:08:24] The Texas Education Agency expects to save $15 to $20 million annually

[00:08:30] by reducing the need for temporary human scores.

[00:08:33] While some educators are concerned about the system, TEA assures safety nets are in place,

[00:08:39] including human rescorings and addressing AI confusion.

[00:08:43] The new scoring system was trained using 3000 exam responses that had already received human grading.

[00:08:51] The FDA has approved an AI-driven test developed by Prenosis to predict the risk of sepsis,

[00:08:58] making it the first algorithmic AI-driven diagnostic tool for sepsis to receive FDA

[00:09:04] authorization.

[00:09:05] The test uses an algorithm trained on over 100,000 blood samples and clinical data

[00:09:11] to classify a patient's risk of sepsis into four categories.

[00:09:15] Other companies such as Epic Systems have also developed AI-driven diagnostics for sepsis,

[00:09:21] but it makes questions about the accuracy of their algorithms.

[00:09:25] A group of graduate students from UC Berkeley has developed an AI-powered tool called

[00:09:30] PolyWatch to detect congressional insider trading.

[00:09:34] The tool analyzes publicly available stock filings, hearing schedules,

[00:09:38] committee assignments and sponsored travel to identify suspicious activities.

[00:09:43] PolyWatch aims to increase transparency and hold public officials accountable for their stock trades.

[00:09:48] Although still in its early stages, the tool has the potential to fundamentally change how

[00:09:53] journalists, researchers, and the general public track and monitor stock trades by members of

[00:09:59] Congress. New York City Mayor Eric Adams is defending the city's AI Chatbot,

[00:10:05] MyCity, despite reports that it has been giving incorrect advice that would lead to breaking

[00:10:10] the law. The Chatbot launched as a pilot program aims to provide business owners with

[00:10:15] accurate information that has been found to be wrong several times.

[00:10:19] The city acknowledges the issues and plans to fix them in collaboration with Microsoft,

[00:10:23] the provider of the AI service. Kaseya has partnered with Win.ai, a sales tech startup to

[00:10:30] enhance sales efficiency through generative AI. Kaseya has licensed Win.ai's real-time AI

[00:10:36] assistant for over a thousand sales professionals resulting in an 80% reduction in call documentation

[00:10:42] time. Win.ai's AI assistant enables hands-free note-taking during customer conversations,

[00:10:49] improving interactions and reducing administrative workloads.

[00:10:54] Why do we care? Sure, IT providers and MSPs are probably dying to hate on Kaseya,

[00:11:00] but it's the use case you care about, the obvious one of helping with documentation.

[00:11:05] Summarization and documentation are strong AI use cases. More revolutionary will be data

[00:11:11] analytics solutions in medicine or politics. And Mayor Adams is a warning about not doing

[00:11:17] your diligence before release. With as many breaches and security concerns as I report in this show,

[00:11:25] it should be obvious that cybersecurity is not just about technology, but also the human expertise

[00:11:31] needed to interpret and respond to complex threats. Huntress is focused on elevating

[00:11:37] SMBs and MSPs around the world. Huntress has a suite of fully-managed cybersecurity solutions

[00:11:44] powered by a 24x7 human-led SOC dedicated to continuous monitoring, expert investigation,

[00:11:50] and rapid response. And the proof is the execution. Huntress is the number one rated EDR for SMBs on

[00:11:59] G2. Want to know more about the platform? Visit huntress.com.mspradio to learn more.

[00:12:07] Thanks for listening! Today is National Pet Day. We hugged dogs yesterday, so today all the pets get

[00:12:15] some attention. And if you're not into that, it's also National Cheese Fondue Day and National

[00:12:20] Eat Track Tape Day. Have a question you want answered? We take those. Ideally,

[00:12:25] his voicemem was our video to question at mspradio.com. I answer list of questions live

[00:12:31] each week on our Wednesday live show on YouTube and LinkedIn. And if you got a thought or

[00:12:35] comment, put it in comments if you're on YouTube or reach out on LinkedIn if you're listening to the

[00:12:39] podcast. Talk to you again tomorrow! The business of tech is written and produced by me, Dave

[00:12:47] Sobel, under ethics guidelines. Post it at businessof.tech. If you like the content,

[00:12:52] please make sure to hit that like button, follow, or subscribe. It's free and easy

[00:12:58] and the best way to support the show and help us grow. You can also check out our Patreon

[00:13:03] where you can join the business of tech community at patreon.com slash mspradio or buy our Why Do We

[00:13:11] Care merch at businessof.tech. Finally, if you're interested in advertising on this show,

[00:13:17] visit mspradio.com slash engage. Once again, thanks for listening to me and I will talk to you

[00:13:24] again on our next episode of the business of tech. Part of the msp radio network.