Showing posts with label AI Regulation. Show all posts
Showing posts with label AI Regulation. Show all posts

Tuesday, December 16, 2025

From Disney to City Hall: New Partnerships, Policies, and Public Impact

When attempting to extrapolate how the present developments in generative AI will lead to the platform of the future, there are a couple of stories to dive deeper into—and what we're looking at this week.

One, which was kind of a small development, was Google's Deep Research. The ability now is that developers are able to use the API to build apps with this capability. If you're not familiar with Deep Research, it's definitely something you should check out. Google was really the first AI offering to provide this functionality in its Gemini model, and it is something quite amazing.

Understanding that the platform of the future is going to be a composite technology—this is really where Deep Research comes in. The LLM will not just respond to prompts but will actually go out and research on the web. It's that combination of using its natural language processing capabilities to actually go do something. This is where we have our first glimpse into what agentic looks like, and it's pretty amazing. 

I demoed Deep Research with other members of the faculty, and they were impressed. It's a good illustration of how this could potentially help tax professionals and accounting professionals do research. Even if you don't trust the output, it does provide a good set of links at the bottom of the page, which enables you to verify. If you're looking at it in-app, there's a way to export Deep Research into a Google Doc. But if you're looking within the actual interface of Gemini, you can go paragraph by paragraph to see what links it's providing and then refine your research from there.

The second story to check out is OpenAI's partnership with Disney. I've felt for a very long time that generative AI will be the ultimate amplifier for storytelling and user-generated content. What it does is give the capability to someone who has great ideas but isn't a professionally trained writer—someone who's not able to get to Hollywood but has great ideas. This is similar to user-generated content, like Dude Perfect—someone who has great ideas but wants to tell a story about Darth Vader. With Disney now owning these kinds of properties—Marvel, Star Wars—it will open up that capability for people with great ideas to create.

Now, I think it's an interesting aspect here, because there's a tendency to think that generative AI is just about cheating. As OpenAI and Disney kind of finalize their partnership, what'll be interesting is to see the tools that are necessary to create the tool of the future. What does that look like when you're trying to create an animation? What does a GenAI video content generator look like? This could be financial salvation for OpenAI. I think the scope is limited right now—it's just social animation, so you're only using Sora within the context of the app.

But regardless, we've seen the success of user-generated content, and it's arguably one of the reasons why Quibi failed during the pandemic—star power doesn't have that much power anymore. That's something of a bygone era. Now, what matters is user-generated content. You can see this with vlogging videos out there that really illustrate the capability of being able to tell a story in a unique way.

There'll be many who argue that this is not real art, that this is not the same as "real" human-generated content—and that's fair. But I would articulate that this is similar to electronic music. People would argue that electronic dance music, or EDM, is not real music. But it created a different genre. It's not like classical music; it's not like rock music. So if you're a fan of CCR—Creedence Clearwater Revival—you're going to argue that techno music is not real music. But it created a different genre and a different type of audience. And I think that either the story is good or bad—that's kind of what it comes down to.

What will enable OpenAI to potentially become its own kind of movie studio is the ability to create a specific filmmaking tool. Most video editors use tools like Premiere Pro or DaVinci Resolve. Most learn these tools through YouTube videos - no certification required.   

And I think that's one of the pathways to the future, because there's been a lot of anti-OpenAI rhetoric out there—comparisons to Myspace and things like that by certain detractors. However, the challenge is to chart the pathway to the future: how do we build something new?

This is where AI builds, not just displaces. The appetite for professionally crafted stories—Star Wars, anime, the next great cinematic experience—isn't going anywhere. But alongside it, we're watching a new genre emerge: stories created by everyday people, powered by tools that didn't exist five years ago. The next Dude Perfect might not just be doing trick shots—they might be producing their own animated series. That's not a threat to storytelling. That's its next chapter.

Disney and OpenAI Strike Landmark Deal to Bring Iconic Characters to Generative AI


The Walt Disney Company and OpenAI announced a three-year licensing and partnership agreement that will allow OpenAI’s generative video platform, Sora, and ChatGPT Images to create fan-inspired short-form videos and images using more than 200 characters from Disney, Pixar, Marvel, and Star Wars. Users will be able to generate short, shareable social videos featuring iconic characters, environments, and props, with curated selections eventually streaming on Disney+. Beyond licensing, Disney will become a major OpenAI customer, integrating OpenAI’s APIs into new products and experiences, including Disney+, and deploying ChatGPT internally. Disney will also make a $1 billion equity investment in OpenAI. Both companies emphasized responsible AI use, including safeguards for creators’ rights and user safety, positioning the agreement as a model for collaboration between AI and entertainment leaders.

(Source: OpenAI)

  • Generative fan content expands: Fans will be able to create short AI-generated videos and images using hundreds of Disney-owned characters.
  • Strategic partnership deepens: Disney will invest $1 billion in OpenAI and adopt its technology across products and internal operations.
  • Responsible AI focus: Both companies stress protections for creators, users, and intellectual property.

How Saskatoon Is Using AI to Keep City Buses—and Services—Running Smoothly

Saskatoon Transit is using artificial intelligence to improve fleet reliability by identifying mechanical issues before buses break down. Hardware installed on more than 130 buses sends real-time sensor data to a central system, where AI analyzes performance and flags maintenance needs. Since launching as a pilot in 2023, the system has reduced unscheduled maintenance, lowered parts costs, and improved service reliability. AI is also being used across Saskatoon’s water services, waste management, administration, and energy efficiency systems. Nationally, adoption is growing, with many Canadian municipalities using or evaluating AI tools to support operations. While cost, privacy, and data accuracy remain concerns, experts say AI is increasingly seen as a way to modernize services without displacing workers.
(Source: CTV News)

  • Predictive maintenance in transit: AI helps Saskatoon detect bus issues early, reducing breakdowns and costs.
  • Municipal adoption is rising: Cities across Canada are experimenting with AI in services like HR, infrastructure, and traffic analysis.
  • Efficiency without layoffs: AI is being used mainly to automate routine tasks rather than replace workers.

The Real AI Fear Isn’t a Bubble—it’s Mass Layoffs and Inequality

A commentary in The Guardian argues that public anxiety around artificial intelligence centers less on speculative tech bubbles and more on the risk of widespread job losses and rising income inequality. Citing warnings from AI executives, economists, and policymakers, the piece highlights concerns that AI could eliminate millions of jobs, particularly entry-level white-collar roles. MIT economist and Nobel laureate Daron Acemoglu describes two possible paths for AI: one that maximizes automation and job cuts, and another that enhances workers’ skills and productivity. The article calls for stronger government intervention, including retraining programs, healthcare reform, shorter workweeks, and expanded unemployment insurance, to ensure AI benefits are more evenly distributed.
(Source: The Guardian)

  • Job security is the main concern: Many fear AI will lead to mass layoffs and greater inequality.
  • Two paths for AI: Experts argue AI can either replace workers or be designed to augment their skills.
  • Policy response needed: Governments may need to act to protect workers and modernize safety nets.

Trump Executive Order Seeks to Block State AI Rules in Favor of National Framework

President Donald Trump signed an executive order aimed at preventing states from enforcing their own artificial intelligence regulations while the federal government works toward a unified national framework. Administration officials say the move is intended to prevent a patchwork of state rules that could slow innovation and weaken US competitiveness. Critics argue the order could undermine consumer protections and accountability, particularly in areas such as deepfakes, discrimination, healthcare, and policing. The decision has exposed divisions within Congress and the Republican Party, and legal experts expect court challenges. Many stakeholders now say Congress faces increased pressure to pass comprehensive federal AI legislation.
(Source: CNN)

  • Federal preemption effort: The executive order seeks to limit state-level AI regulation.
  • Ongoing debate: Supporters cite innovation and competitiveness, while critics warn of weakened safeguards.
  • Legislative pressure grows: Congress may need to establish clear federal AI rules.

Google and OpenAI Trade Blows as Deep Research and GPT-5.2 Launch Side by Side

Google unveiled a major upgrade to its Gemini Deep Research agent on the same day OpenAI released GPT-5.2, highlighting intensifying competition in advanced AI. Built on Gemini 3 Pro, the new agent allows developers to embed deep research capabilities into their own applications through a new Interactions API. Google says the tool is designed to handle large volumes of information while minimizing hallucinations during complex, multi-step tasks. The company introduced a new open-source benchmark to demonstrate progress, though OpenAI’s near-simultaneous release of GPT-5.2 quickly shifted attention back to the broader AI rivalry.
(Source: TechCrunch)

  • More capable research agents: Google’s update enables deeper, more autonomous research workflows.
  • Accuracy remains critical: Reducing hallucinations is key for long-running AI tasks.
  • Competition is accelerating: Major AI players continue to release upgrades at a rapid pace.
Author: Malik D. CPA, CA, CISA. The opinions expressed here do not necessarily represent UWCISA, UW,  or anyone else. This post was written with the assistance of an AI language model. 

Monday, March 18, 2024

Five Top Tech Takeaways: AI Agents take on Software Engineering, Grok Open-Sourced, Figure's OpenAI Assisted Robot, TikTok Ban, and EU's AI Legislation

Robot Developer who Takes out the Trash

x.ai Goes Public: Musk's Open Sources Grok

Elon Musk's xAI has made a significant move in the AI landscape by open-sourcing its AI chatbot Grok on GitHub, enabling researchers and developers to build upon and influence its future iterations. This move is part of a broader trend of AI democratization and competition among tech giants such as OpenAI, Meta, and Google. Grok, described as a "314 billion parameter Mixture-of-Experts model," offers a base model for various applications without being fine-tuned for specific tasks. While the release under the Apache 2.0 license permits commercial use, it notably excludes the training data and real-time data connections. This strategy aligns with Musk's advocacy for open-source AI, contrasting with the practices of some firms that maintain proprietary models or offer limited open-source access. The initiative reflects a larger dialogue on openness and accessibility in AI development, with potential implications for innovation and the direction of future AI technologies. 

Key Takeaways:
  • Elon Musk's xAI has open-sourced its AI chatbot Grok, aiming to foster innovation and competition in the AI sector.
  • Grok is released as a versatile, yet unrefined model under the Apache 2.0 license, emphasizing commercial use without offering training data or real-time data connections.
  • Musk's approach to open-sourcing contrasts with other tech giants, highlighting a broader industry debate on the balance between proprietary and open-source AI models.
(Source: The Verge)

Navigating the EU's AI Act: Implications for Consumers and Tech Giants

The European Union's proposed AI law, recently endorsed by the European Parliament, represents a significant step toward regulating AI technologies to ensure consumer safety and trust. Set to become law within weeks, it introduces comprehensive measures to regulate AI, including stringent definitions, prohibited practices, and special provisions for high-risk systems. The law aims to foster a safer AI environment, with mandatory vetting and safety protocols akin to those used in banking apps. It addresses concerns over AI misuse, including manipulative systems, social scoring, and unauthorized biometric categorization, while exempting military, defense, and national security applications. For high-risk applications, such as those in critical infrastructure, healthcare, and education, the law mandates accuracy, risk assessments, human oversight, and transparency. Additionally, it tackles the complexities of generative AI and deepfakes, requiring disclosure and adherence to copyright laws. Despite mixed reactions from tech giants, the EU's pioneering legislation could significantly influence global AI regulation standards, ensuring AI's responsible development and use. 

The article also noted the fines that can be imposed under the legislation:
"Fines will range from €7.5m or 1.5% of a company’s total worldwide turnover – whichever is higher – for giving incorrect information to regulators, to €15m or 3% of worldwide turnover for breaching certain provisions of the act, such as transparency obligations, to €35m, or 7% of turnover, for deploying or developing banned AI tools. There will be more proportionate fines for smaller companies and startups."

(Source: The Guardian)

Key Takeaways:
  • The EU's AI regulation marks a crucial advance in AI governance, emphasizing consumer safety and the responsible use of AI technologies.
  • It categorically bans or regulates AI applications based on risk levels, from manipulative technologies to high-risk systems in vital sectors, ensuring oversight and transparency.
  • The legislation's impact extends beyond the EU, setting a precedent for global AI practices, amid tech industry concerns over innovation constraints and regulatory burdens.
TikTok Under Fire: National Security Concerns Prompt Legislative Action

The U.S. Congress has made significant progress toward imposing restrictions on TikTok, a move with potential widespread effects on social media within the nation. The House of Representatives passed the "Protecting Americans from Foreign Adversary Controlled Applications Act," aimed at TikTok and other apps owned by countries considered foreign adversaries, including China. The bill mandates that TikTok's Chinese owner, ByteDance, must either sell the platform within 180 days or face a ban in the U.S. This legislation reflects broader concerns over national security and the influence of foreign powers on American digital platforms. Despite the overwhelming support in the House, the bill's future in the Senate remains uncertain, as it competes with other legislative priorities.

Key takeaways:
  • The U.S. House of Representatives has passed a bill potentially leading to a TikTok ban unless its Chinese owners divest, signaling heightened scrutiny on foreign-controlled social media.
  • Concerns over national security and the influence of foreign adversaries are central to the legislative move against TikTok, reflecting broader geopolitical tensions.
  • While the bill has gained significant bipartisan support in the House, its passage in the Senate is not assured, underscoring the complexities of legislative action on social media regulation
(Source: CBC)

    The Dawn of Devin: Autonomous AI Takes Software Engineering to New Heights 
    Cognition AI's release of an AI program named Devin, which performs tasks typically done by software engineers, has sparked excitement and concern in the tech industry. Devin is capable of planning, coding, testing, and implementing solutions, showcasing a significant advancement beyond what chatbots like ChatGPT and Gemini offer. This development represents a growing trend towards AI agents that can take actions to solve problems independently, a departure from merely generating text or advice. Although impressive, these AI agents, including Google DeepMind's SIMA, which can play video games with considerable skill, still face challenges related to error rates and potential failures. However, the ongoing refinement and potential applications of these AI agents in various fields hint at a future where they could dramatically change how tasks are approached and completed.

    Key takeaways:
  • Devin, an AI developed by Cognition AI, demonstrates advanced capabilities in software development, challenging traditional roles within the tech industry.
  • The emergence of AI agents capable of independently solving problems signifies a significant evolution from earlier AI models focused on generating responses or performing predefined tasks.
  • Despite their potential, these AI agents still face challenges in accuracy and reliability, highlighting the need for continued development to minimize errors and their consequences
(Source: WIRED)

In the following video, Cognition AI, demonstrates how Devin can perform a job posted on Upwork:


Meet Figure 01: The Humanoid Robot That Converses and Multitasks

Figure, an AI robotics developer, recently unveiled its first humanoid robot, Figure 01, showcasing its ability to engage in real-time conversations and perform tasks simultaneously using generative AI from OpenAI. This collaboration enhances the robot's visual and language intelligence, allowing for swift and precise actions. In a demo, Figure 01 demonstrated its multitasking prowess by identifying objects and handling tasks in a kitchen setup, fueled by its capacity to describe its visual experiences, plan, and execute actions based on a multimodal AI model. This model integrates visual data and speech, enabling the robot to respond to verbal commands and interact naturally. The development signifies a leap forward in AI and robotics, merging sophisticated AI models with physical robotic bodies, aiming to fulfill practical and utilitarian objectives in various sectors, including space exploration.

Key takeaways:
  • Figure's humanoid robot, Figure 01, can converse and perform tasks in real-time, powered by OpenAI's generative AI technology.
  • The robot's AI integrates visual and auditory data, allowing it to plan actions and respond to commands intelligently.
  • Figure 01's development marks significant progress in combining AI with robotics, potentially revolutionizing practical applications in multiple fields. 
(Source: Decrypt)

Here is the official video from the company, Figure:


Author: Malik Datardina, CPA, CA, CISA. Malik works at Auvenir as a GRC Strategist who is working to transform the engagement experience for accounting firms and their clients. The opinions expressed here do not necessarily represent UWCISA, UW, Auvenir (or its affiliates), CPA Canada or anyone else. This post was written with the assistance of an AI language model. The model provided suggestions and completions to help me write, but the final content and opinions are my own.

Tuesday, June 20, 2023

The Furious Five for June 20: AI, Tech, and Business Stories You May Have Missed

Redditor Robots Rioting
McKinsey's Take on the Trillion-Dollar Potential of Generative AI

Generative AI's potential to significantly boost global economic value could add between $2.6 trillion and $4.4 trillion annually across diverse use cases, enhancing AI's overall impact by 15-40%. This technology can automate up to 70% of current work activities and may advance the timeline for automation of half of today's tasks to between 2030 and 2060. Further, it could stimulate labour productivity growth between 0.1% to 0.6% per annum through 2040. However, fully leveraging generative AI's benefits necessitates managing inherent risks and supporting workforce transitions with new skills development. Despite its potential, the journey toward full realization is nascent, and substantial challenges remain. (Source: McKinsey)

Revolt at Reddit: Reddit's API Pricing Protest

The "Reddit Revolt" is a significant user and developer backlash against Reddit due to changes in its API pricing, which are predicted to lead to the shutdown of many third-party apps like Apollo and "rif is fun for Reddit". The discontent has been fuelled by perceived miscommunications and alleged dishonesty from Reddit's CEO, Steve Huffman, regarding these changes, including incoming restrictions on third-party apps displaying NSFW content. In response, over 100 subreddits have "gone dark" in protest, and nearly 4,500 communities are pledged to do so. Some speculate that these changes are due to Reddit's plans to go public. (Source: TheVerge)

Mistral AI Sets Sights on OpenAI with $113 Million Seed Funding

Mistral AI, a month-old startup based in Paris, has raised a staggering $113 million in a seed funding round to develop large language models and generative AI, aiming to compete with OpenAI. The funding round was led by Lightspeed Venture Partners and involved significant contributors from across Europe, including Bpifrance and former Google CEO Eric Schmidt. Mistral AI, co-founded by former Google DeepMind and Meta professionals, plans to focus on open-source solutions and target enterprises. The company's first models for text-based generative AI are expected to be launched in 2024. It aims to provide businesses with tools that simplify the integration and use of AI, with a particular focus on open-source models that use publicly available data. This strategy is designed to avoid legal issues related to training data. (Source: TechCrunch)

First-of-its-kind AI Legislation Advances in EU Parliament

The European Union's parliament is advancing draft legislation known as the AI Act, potentially the first comprehensive set of regulations governing Artificial Intelligence (AI) in the West. This act includes prohibitions on real-time remote biometric surveillance in public spaces and restrictions on developing facial-recognition databases from surveillance footage or internet scraping. The legislation also aims to control predictive policing systems and regulate how companies train AI models with large datasets. In certain scenarios, companies would have to disclose when content is AI-generated and ensure their AI models don't produce illegal content. Businesses would also need to summarize the copyrighted data used to train their models, giving content creators a potential claim to a share of profits. The proposed act could impose fines of up to 6-7% of a company’s global revenue for certain non-compliance cases. Negotiations to finalize the legislation are to begin immediately, with the aim of reaching a deal before year-end. Some tech researchers support these rules, arguing they could help establish safety standards and slow down the rush to release advanced AI tools. However, tech companies and lobbyists argue that overly prescriptive rules could hinder innovation. (Source: WSJ)

AMD's Chase for Trillions Stumbles: Stock Prices Slide Despite Bold AI Chip Announcement

AMD revealed details of its upcoming artificial intelligence (AI) chip designed to compete with industry leader Nvidia. Expected to be available in Q3 with mass production in Q4, the chip boasts 192 gigabytes of memory, potentially assisting tech companies in managing the costs of AI services like ChatGPT. However, AMD refrained from disclosing the early adopters or the price of the new chip. While the lack of a major client announcement led to a dip in AMD's share price, the company remains optimistic about the chip's potential. The firm also announced high-volume shipments of a general-purpose central processor chip named "Bergamo" to companies, including Meta Platforms. AMD's strategic moves indicate their intention to vie for market share in the AI chip market, currently dominated by Nvidia, and bolster their valuation. (Source: Reuters)

Author: Malik Datardina, CPA, CA, CISA. Malik works at Auvenir as a GRC Strategist that is working to transform the engagement experience for accounting firms and their clients. The opinions expressed here do not necessarily represent UWCISA, UW, Auvenir (or its affiliates), CPA Canada or anyone else. This post was written with the assistance of an AI language model. The model provided suggestions and completions to help me write, but the final content and opinions are my own.